title
TensorFlow Object Detection | Realtime Object Detection with TensorFlow | TensorFlow Python |Edureka
description
π₯ AI & Deep Learning Using TensorFlow (Use Code "πππππππππ") - https://www.edureka.co/ai-deep-learning-with-tensorflow
This Edureka video will provide you with a detailed and comprehensive knowledge of TensorFlow Object detection and how it works. It will also provide you with the details on how to use Tensorflow to detect objects in deep learning method. Below are the topics covered in this tutorial:
1. What is Object Detection?
2. Industrial use of Object Detection
3. Object Detection Workflow
4. What is Tensorflow?
5. Object Detection using Tensorflow - Demo
6. Live Object Detection using Tensorflow- Demo
Subscribe to our channel to get video updates. Hit the subscribe button above.
Check our complete Deep Learning With TensorFlow playlist here: https://goo.gl/cck4hE
PG in Artificial Intelligence and Machine Learning with NIT Warangal : https://www.edureka.co/post-graduate/machine-learning-and-ai
Post Graduate Certification in Data Science with IIT Guwahati - https://www.edureka.co/post-graduate/data-science-program
(450+ Hrs || 9 Months || 20+ Projects & 100+ Case studies)
- - - - - - - - - - - - - -
How it Works?
1. This is 21 hrs of Online Live Instructor-led course. Weekend class: 7 sessions of 3 hours each.
2. We have a 24x7 One-on-One LIVE Technical Support to help you with any problems you might face or any clarifications you may require during the course.
3. At the end of the training you will have to undergo a 2-hour LIVE Practical Exam based on which we will provide you a Grade and a Verifiable Certificate!
- - - - - - - - - - - - - -
About the Course
Edureka's Deep learning with Tensorflow course will help you to learn the basic concepts of TensorFlow, the main functions, operations and the execution pipeline. Starting with a simple βHello Wordβ example, throughout the course you will be able to see how TensorFlow can be used in curve fitting, regression, classification and minimization of error functions. This concept is then explored in the Deep Learning world. You will evaluate the common, and not so common, deep neural networks and see how these can be exploited in the real world with complex raw data using TensorFlow. In addition, you will learn how to apply TensorFlow for backpropagation to tune the weights and biases while the Neural Networks are being trained. Finally, the course covers different types of Deep Architectures, such as Convolutional Networks, Recurrent Networks and Autoencoders.
Delve into neural networks, implement Deep Learning algorithms, and explore layers of data abstraction with the help of this Deep Learning with TensorFlow course.
- - - - - - - - - - - - - -
Who should go for this course?
The following professionals can go for this course:
1. Developers aspiring to be a 'Data Scientist'
2. Analytics Managers who are leading a team of analysts
3. Business Analysts who want to understand Deep Learning (ML) Techniques
4. Information Architects who want to gain expertise in Predictive Analytics
5. Professionals who want to captivate and analyze Big Data
6. Analysts wanting to understand Data Science methodologies
However, Deep learning is not just focused to one particular industry or skill set, it can be used by anyone to enhance their portfolio.
- - - - - - - - - - - - - -
Why Learn Deep Learning With TensorFlow?
TensorFlow is one of the best libraries to implement Deep Learning. TensorFlow is a software library for numerical computation of mathematical expressions, using data flow graphs. Nodes in the graph represent mathematical operations, while the edges represent the multidimensional data arrays (tensors) that flow between them. It was created by Google and tailored for Machine Learning. In fact, it is being widely used to develop solutions with Deep Learning.
Machine learning is one of the fastest-growing and most exciting fields out there, and Deep Learning represents its true bleeding edge. Deep learning is primarily a study of multi-layered neural networks, spanning over a vast range of model architectures. Traditional neural networks relied on shallow nets, composed of one input, one hidden layer and one output layer. Deep-learning networks are distinguished from these ordinary neural networks having more hidden layers, or so-called more depth. These kinds of nets are capable of discovering hidden structures within unlabeled and unstructured data (i.e. images, sound, and text), which constitutes the vast majority of data in the world.
For more information, please write back to us at sales@edureka.co or call us at IND: 9606058406 / US: 18338555775 (toll-free).
Facebook: https://www.facebook.com/edurekaIN/
Twitter: https://twitter.com/edurekain
LinkedIn: https://www.linkedin.com/company/edureka
detail
{'title': 'TensorFlow Object Detection | Realtime Object Detection with TensorFlow | TensorFlow Python |Edureka', 'heatmap': [{'end': 689.451, 'start': 572.266, 'weight': 0.754}, {'end': 754.58, 'start': 718.763, 'weight': 0.782}], 'summary': 'Covers the significance of object detection in computer vision, industrial applications, feature extraction, tensorflow in deep learning, and tensorflow object detection, including methods achieving 94% and 92% accuracy.', 'chapters': [{'end': 102.288, 'segs': [{'end': 64.45, 'src': 'embed', 'start': 7.045, 'weight': 0, 'content': [{'end': 16.754, 'text': 'Creating accurate machine learning models capable of localizing and identifying multiple objects in a single image remained a core challenge in computer vision,', 'start': 7.045, 'duration': 9.709}, {'end': 21.92, 'text': 'but with recent advancement in deep learning and deep learning based computer vision models,', 'start': 16.754, 'duration': 5.166}, {'end': 25.343, 'text': 'object detection applications are easier to develop than ever before.', 'start': 21.92, 'duration': 3.423}, {'end': 31.373, 'text': 'So guys I Kislev from Edreka welcome you to this session on object detection using tensorflow.', 'start': 26.207, 'duration': 5.166}, {'end': 35.036, 'text': "So before moving forward, let's have a quick look at the agenda of this session.", 'start': 31.853, 'duration': 3.183}, {'end': 38.46, 'text': "So first of all, I'll be starting with explaining you guys.", 'start': 35.537, 'duration': 2.923}, {'end': 40.883, 'text': 'What is exactly object detection?', 'start': 38.52, 'duration': 2.363}, {'end': 48.731, 'text': "then we'll have a look at the object detection workflow, how exactly it works and the various industrial use cases of object detection.", 'start': 40.883, 'duration': 7.848}, {'end': 56.749, 'text': "Then we'll understand what is tensorflow and I'll end this video with two demos, the first one being the tensorflow object detection,", 'start': 49.286, 'duration': 7.463}, {'end': 64.45, 'text': 'where we can detect objects using our own input images, and the second demo is based on the live object detection,', 'start': 56.749, 'duration': 7.701}], 'summary': 'Advancements in deep learning have made object detection applications easier to develop, as explained in a session on object detection using tensorflow.', 'duration': 57.405, 'max_score': 7.045, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls7045.jpg'}], 'start': 7.045, 'title': 'Object detection with tensorflow', 'summary': 'Discusses the significance of object detection in computer vision, the workflow of object detection, industrial use cases, and the impact of deep learning in improving real-time object detection performance.', 'chapters': [{'end': 102.288, 'start': 7.045, 'title': 'Object detection with tensorflow', 'summary': 'Discusses the significance of object detection in computer vision, the workflow of object detection, the industrial use cases, and the impact of deep learning in improving the performance of real-time object detection.', 'duration': 95.243, 'highlights': ['Object detection allows for the recognition, detection, and localization of multiple objects within an image, providing a better understanding of the image as a whole.', 'Recent advancements in deep learning have significantly improved the performance of object detection, enabling real-time use cases.', 'The chapter covers the agenda of the session, including explaining object detection, its workflow, industrial use cases, and demonstrations of tensorflow object detection and live object detection.', 'Object detection is the process of finding instances of real-world objects in images or videos, such as faces, bicycles, and buildings.']}], 'duration': 95.243, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls7045.jpg', 'highlights': ['Recent advancements in deep learning have significantly improved the performance of object detection, enabling real-time use cases.', 'Object detection allows for the recognition, detection, and localization of multiple objects within an image, providing a better understanding of the image as a whole.', 'The chapter covers the agenda of the session, including explaining object detection, its workflow, industrial use cases, and demonstrations of tensorflow object detection and live object detection.', 'Object detection is the process of finding instances of real-world objects in images or videos, such as faces, bicycles, and buildings.']}, {'end': 384.914, 'segs': [{'end': 151.189, 'src': 'embed', 'start': 124.802, 'weight': 8, 'content': [{'end': 132.048, 'text': 'Now the first use case is the facial recognition the popular application include face detection and people counting.', 'start': 124.802, 'duration': 7.246}, {'end': 138.374, 'text': 'Have you ever noticed how Facebook detects your face when you upload a photo that is due to their facial recognition algorithm.', 'start': 132.449, 'duration': 5.925}, {'end': 142.497, 'text': 'Now, this is a simple example of object detection that we can see in our daily life.', 'start': 138.854, 'duration': 3.643}, {'end': 147.362, 'text': 'Even the DSLR cameras and our phone cameras also detect the faces.', 'start': 143.218, 'duration': 4.144}, {'end': 151.189, 'text': 'Object detection can also be used for people counting.', 'start': 148.346, 'duration': 2.843}], 'summary': 'Facial recognition and object detection used for people counting in popular applications like facebook and cameras.', 'duration': 26.387, 'max_score': 124.802, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls124802.jpg'}, {'end': 215.808, 'src': 'embed', 'start': 172.496, 'weight': 0, 'content': [{'end': 180.5, 'text': 'finding a specific object through visual inspection is a basic task that is involved in multiple industrial processes like sorting,', 'start': 172.496, 'duration': 8.004}, {'end': 185.503, 'text': 'inventory management, machining, quality management, packaging and much more.', 'start': 180.5, 'duration': 5.003}, {'end': 189.985, 'text': 'inventory management can be very tricky, as items are hard to track in real time.', 'start': 185.503, 'duration': 4.482}, {'end': 193.407, 'text': 'Something is always added removed and moved every day.', 'start': 190.425, 'duration': 2.982}, {'end': 200.536, 'text': 'Systems can perform automatic object counting and localization that will allow you to improve inventory accuracy.', 'start': 193.911, 'duration': 6.625}, {'end': 208.002, 'text': 'Now, this is one of the most recent and I should say one of the most exciting use cases of object detection, which is the self-driving car.', 'start': 200.997, 'duration': 7.005}, {'end': 213.887, 'text': 'It is a vehicle that is capable of sensing its environment and navigating without human input.', 'start': 208.483, 'duration': 5.404}, {'end': 215.808, 'text': 'Now, how is it done now?', 'start': 214.307, 'duration': 1.501}], 'summary': 'Visual inspection is crucial in industrial processes like inventory management, where automatic object counting can improve accuracy.', 'duration': 43.312, 'max_score': 172.496, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls172496.jpg'}, {'end': 306.64, 'src': 'embed', 'start': 241.997, 'weight': 2, 'content': [{'end': 246.819, 'text': 'Now our final use case is the most important one, which is the security.', 'start': 241.997, 'duration': 4.822}, {'end': 254.201, 'text': 'now it is used in the banking industry as well as the phone industry, or I should say the technologically industry, a lot.', 'start': 246.819, 'duration': 7.382}, {'end': 261.523, 'text': 'So it is using the banking industry to identify fraud the forgery of notes or the currency and also theft.', 'start': 254.721, 'duration': 6.802}, {'end': 268.405, 'text': 'Now, one of the most important use of object action is the facial recognition used in the smartphones nowadays,', 'start': 262.383, 'duration': 6.022}, {'end': 270.346, 'text': 'which is popularly known as face unlock.', 'start': 268.405, 'duration': 1.941}, {'end': 275.042, 'text': 'Now that we have understood what are the various industrial use cases of object detection.', 'start': 271.039, 'duration': 4.003}, {'end': 282.388, 'text': "Let's have a look at the object detection workflow not typically there are three steps involved in object detection workflow.", 'start': 275.663, 'duration': 6.725}, {'end': 287.712, 'text': 'First of all, we have the training data which we input and create our model.', 'start': 282.868, 'duration': 4.844}, {'end': 297.36, 'text': 'Now this model is created using the visual features which are extracted from the images or the digits or whatever be your input of the training data.', 'start': 288.733, 'duration': 8.627}, {'end': 304.519, 'text': 'Now these are evaluated and is determined whether which objects are present in the proposal based on the visual features.', 'start': 297.975, 'duration': 6.544}, {'end': 306.64, 'text': 'now, upon extracting these features', 'start': 304.519, 'duration': 2.121}], 'summary': 'Object detection is used in banking and phone industries for security, including fraud detection and facial recognition for smartphones.', 'duration': 64.643, 'max_score': 241.997, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls241997.jpg'}, {'end': 355.607, 'src': 'embed', 'start': 328.775, 'weight': 5, 'content': [{'end': 335.378, 'text': 'Now there is an important trade-off being made here, which is between the number of regions versus the computational complexity.', 'start': 328.775, 'duration': 6.603}, {'end': 338.379, 'text': 'We learn more about these later in this video.', 'start': 335.858, 'duration': 2.521}, {'end': 341.521, 'text': 'Now in object detection frameworks.', 'start': 339.36, 'duration': 2.161}, {'end': 349.264, 'text': 'people typically use pre-trained image classification models to extract visual features, as these tend to generalize fairly well.', 'start': 341.521, 'duration': 7.743}, {'end': 355.607, 'text': 'The goal of feature extraction is to reduce a variable sized image to a fixed set of visual features.', 'start': 349.924, 'duration': 5.683}], 'summary': 'Trade-off between regions and computational complexity in object detection frameworks is crucial.', 'duration': 26.832, 'max_score': 328.775, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls328775.jpg'}, {'end': 392.851, 'src': 'embed', 'start': 370.326, 'weight': 6, 'content': [{'end': 384.914, 'text': 'They all have the exact same objective extracting features from the input image that are representative for the task at hands and use these features to determine the class of the image now that we have understood the basic workflow of an object detection algorithm.', 'start': 370.326, 'duration': 14.588}, {'end': 392.851, 'text': "Let's go ahead and understand what is tensorflow now tensorflow as you can see is divided into two parts which are the tensor and flow.", 'start': 385.307, 'duration': 7.544}], 'summary': 'Object detection algorithm aims to extract image features and determine image class using tensorflow divided into tensor and flow.', 'duration': 22.525, 'max_score': 370.326, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls370326.jpg'}], 'start': 102.288, 'title': 'Industrial applications and feature extraction in object detection', 'summary': 'Covers industrial applications of object detection, including facial recognition for people counting, industrial processes, inventory management, and self-driving cars, with an emphasis on its role in security and banking industries. it also explains feature extraction in object detection, discussing the workflow of the algorithm, training data input, model creation, visual feature extraction, and the use of pre-trained image classification models, while considering the trade-off between the number of regions and computational complexity.', 'chapters': [{'end': 282.388, 'start': 102.288, 'title': 'Industrial applications of object detection', 'summary': 'Explores the industrial applications of object detection, including facial recognition for people counting, industrial processes, inventory management, and self-driving cars, emphasizing its role in security and banking industries. it highlights the increasing use of object detection in various industries and its impact on tasks such as inventory management and self-driving cars.', 'duration': 180.1, 'highlights': ['Object detection is used in various industrial applications such as facial recognition for people counting, industrial processes, inventory management, and self-driving cars, emphasizing its role in security and banking industries.', 'Facial recognition is a popular application of object detection, used for tasks like face detection, people counting, and face unlock on smartphones, showcasing its widespread use and impact on daily life.', 'Inventory management can be challenging due to the constant movement of items, but object detection allows for automatic counting and localization, improving inventory accuracy and efficiency.', 'Self-driving cars utilize object detection in combination with radar, laser light, GPS, and computer vision to sense their environment and navigate, demonstrating the advanced capabilities and potential impact of object detection technology.', 'Object detection is used in security and banking industries to identify fraud, forgery, and theft, highlighting its crucial role in ensuring security and preventing financial crimes.']}, {'end': 384.914, 'start': 282.868, 'title': 'Feature extraction in object detection', 'summary': 'Explains the workflow of object detection algorithm, including training data input, model creation, visual feature extraction, and the use of pre-trained image classification models. it also discusses the trade-off between the number of regions and computational complexity.', 'duration': 102.046, 'highlights': ['The model is created using visual features extracted from the training data, and these features are evaluated to determine the presence of objects in the proposal.', 'In object detection frameworks, pre-trained image classification models are typically used to extract visual features, which tend to generalize fairly well.', 'Feature extraction aims to reduce a variable-sized image to a fixed set of visual features, typically constructed using strong visual feature extraction methods.', 'There is a trade-off between the number of regions and the computational complexity in the object detection process.']}], 'duration': 282.626, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls102288.jpg', 'highlights': ['Object detection is used in various industrial applications such as facial recognition for people counting, industrial processes, inventory management, and self-driving cars, emphasizing its role in security and banking industries.', 'Facial recognition is a popular application of object detection, used for tasks like face detection, people counting, and face unlock on smartphones, showcasing its widespread use and impact on daily life.', 'Self-driving cars utilize object detection in combination with radar, laser light, GPS, and computer vision to sense their environment and navigate, demonstrating the advanced capabilities and potential impact of object detection technology.', 'Inventory management can be challenging due to the constant movement of items, but object detection allows for automatic counting and localization, improving inventory accuracy and efficiency.', 'Object detection is used in security and banking industries to identify fraud, forgery, and theft, highlighting its crucial role in ensuring security and preventing financial crimes.', 'The model is created using visual features extracted from the training data, and these features are evaluated to determine the presence of objects in the proposal.', 'In object detection frameworks, pre-trained image classification models are typically used to extract visual features, which tend to generalize fairly well.', 'Feature extraction aims to reduce a variable-sized image to a fixed set of visual features, typically constructed using strong visual feature extraction methods.', 'There is a trade-off between the number of regions and the computational complexity in the object detection process.']}, {'end': 896.979, 'segs': [{'end': 467.397, 'src': 'embed', 'start': 430.48, 'weight': 0, 'content': [{'end': 434.101, 'text': 'Now these dimension are not limited to 2 it can be multi-dimensional.', 'start': 430.48, 'duration': 3.621}, {'end': 438.194, 'text': 'Now in TensorFlow the computation is approached as a data flow graph.', 'start': 434.671, 'duration': 3.523}, {'end': 449.843, 'text': 'TensorFlow can run on multiple CPUs and GPUs with optional CUDA and SYCL extensions for general-purpose computing on graphics processing unit and TensorFlow.', 'start': 438.995, 'duration': 10.848}, {'end': 458.71, 'text': 'its flexible architecture allows for easy deployment of computation across a variety of platforms, which are the CPUs, GPUs and the TPUs,', 'start': 449.843, 'duration': 8.867}, {'end': 462.874, 'text': 'and from desktops to clusters of servers, to mobile and edge devices.', 'start': 458.71, 'duration': 4.164}, {'end': 467.397, 'text': 'Now TensorFlow computations are expressed as stateful data flow graphs.', 'start': 463.453, 'duration': 3.944}], 'summary': 'Tensorflow supports multi-dimensional computation on various platforms including cpus, gpus, and tpus.', 'duration': 36.917, 'max_score': 430.48, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls430480.jpg'}, {'end': 539.085, 'src': 'embed', 'start': 513.553, 'weight': 2, 'content': [{'end': 518.836, 'text': 'and using this model we get our final output in which we have the objects detected in an image.', 'start': 513.553, 'duration': 5.283}, {'end': 526.999, 'text': 'Now the image are converted into a numpy array in the tensorflow object detection so that the computation can be made easy.', 'start': 519.576, 'duration': 7.423}, {'end': 535.664, 'text': 'We also use the TF record which is the tensorflow record which contains the record of the image along with the tags such as as you can see here.', 'start': 527.4, 'duration': 8.264}, {'end': 536.904, 'text': 'We have the person tag.', 'start': 535.724, 'duration': 1.18}, {'end': 539.085, 'text': 'We have the dog tag and the horse tag.', 'start': 537.244, 'duration': 1.841}], 'summary': 'Using tensorflow object detection, we identified persons, dogs, and horses in images.', 'duration': 25.532, 'max_score': 513.553, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls513553.jpg'}, {'end': 689.451, 'src': 'heatmap', 'start': 572.266, 'weight': 0.754, 'content': [{'end': 577.909, 'text': 'So it is ideal now next we need to do is to clone the GitHub repository of tensorflow.', 'start': 572.266, 'duration': 5.643}, {'end': 588.014, 'text': 'So for that just go to GitHub and type tensorflow, which is the official GitHub repository of tensorflow, and inside that we have the model section.', 'start': 578.509, 'duration': 9.505}, {'end': 589.115, 'text': 'just go to this models', 'start': 588.014, 'duration': 1.101}, {'end': 593.776, 'text': 'you can either clone this tensorflow model or download it as per your wish.', 'start': 589.892, 'duration': 3.884}, {'end': 596.96, 'text': 'So I have already downloaded the tensorflow model.', 'start': 594.517, 'duration': 2.443}, {'end': 606.05, 'text': 'Now. the tensorflow object detection model uses protobuf to configure model and the training parameters.', 'start': 598.461, 'duration': 7.589}, {'end': 611.85, 'text': 'before the framework can be used, the protobuf libraries must be compiled Now to download protobuf.', 'start': 606.05, 'duration': 5.8}, {'end': 619.697, 'text': 'All you need to do is go to Google slash protobuf in GitHub and here you will have all the different versions of protobuf.', 'start': 612.09, 'duration': 7.607}, {'end': 630.185, 'text': 'So according to your OS, which is Linux Mac OS or the Windows OS or if you are using only the python you can download the required protobuf.', 'start': 620.097, 'duration': 10.088}, {'end': 638.552, 'text': 'So, once you have downloaded tensorflow and protobuf, create a folder in just C, which is known as tensorflow,', 'start': 632.027, 'duration': 6.525}, {'end': 641.862, 'text': 'and in this you will have the models master.', 'start': 639.241, 'duration': 2.621}, {'end': 647.724, 'text': 'extract this and rename it as models and extract the protobuf.', 'start': 641.862, 'duration': 5.862}, {'end': 651.286, 'text': 'now, inside protobuf, you have the bin folder.', 'start': 647.724, 'duration': 3.562}, {'end': 654.147, 'text': 'now all you need to do is go to this bin folder.', 'start': 651.286, 'duration': 2.861}, {'end': 657.548, 'text': 'So let me just open the command prompt here.', 'start': 654.867, 'duration': 2.681}, {'end': 661.109, 'text': "I'm using the anaconda prompt, but you can use the command prompt as well.", 'start': 657.588, 'duration': 3.521}, {'end': 665.631, 'text': 'So once you have downloaded and renamed the models master as models.', 'start': 662.11, 'duration': 3.521}, {'end': 669.76, 'text': 'Go back to the GitHub repository and inside models.', 'start': 666.698, 'duration': 3.062}, {'end': 672.902, 'text': 'You have the research and inside research.', 'start': 669.94, 'duration': 2.962}, {'end': 676.584, 'text': 'There is the object detection model, which we are interested in.', 'start': 672.962, 'duration': 3.622}, {'end': 679.205, 'text': "So let's go to the object detection model here.", 'start': 677.024, 'duration': 2.181}, {'end': 689.451, 'text': 'Now, as you can see, this tensorflow object detection API gives an accurate machine learning model description of how the objects are detected,', 'start': 680.686, 'duration': 8.765}], 'summary': 'To set up the tensorflow object detection model, clone the github repository, download tensorflow model, and protobuf, and configure using protobuf libraries.', 'duration': 117.185, 'max_score': 572.266, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls572266.jpg'}, {'end': 754.58, 'src': 'heatmap', 'start': 718.763, 'weight': 0.782, 'content': [{'end': 725.427, 'text': 'And next, what we need to do is, once you have downloaded and extracted the protobuf, you need to copy this command.', 'start': 718.763, 'duration': 6.664}, {'end': 739.396, 'text': 'Now then you need to go into the tensorflow then you need to go into the models and then inside that you need to go into the research.', 'start': 731.451, 'duration': 7.945}, {'end': 746.621, 'text': 'Now once you are inside the research what you need to do is copy this command and paste it and run this command here.', 'start': 740.157, 'duration': 6.464}, {'end': 754.58, 'text': "So what it'll do is I'll explain here is that it will take all the object.", 'start': 748.419, 'duration': 6.161}], 'summary': 'To set up protobuf, navigate to tensorflow/models/research and run a specific command.', 'duration': 35.817, 'max_score': 718.763, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls718763.jpg'}, {'end': 903.451, 'src': 'embed', 'start': 875.066, 'weight': 4, 'content': [{'end': 881.37, 'text': 'now, an important thing to consider here while selecting a model is that it depends on your system which model you should use.', 'start': 875.066, 'duration': 6.304}, {'end': 890.335, 'text': 'suppose, if your system is low on GPU but has higher RAM, you can go for a model which has a higher speed and a higher MAP point.', 'start': 881.37, 'duration': 8.965}, {'end': 896.979, 'text': 'Now this value should always be high if you are looking for a more accurate prediction in your images.', 'start': 891.516, 'duration': 5.463}, {'end': 903.451, 'text': 'So once all your dependencies are downloaded and you have installed tensorflow and protobuf.', 'start': 897.987, 'duration': 5.464}], 'summary': 'Select a model based on system specs for accurate predictions. consider speed and map point for high ram, low gpu systems.', 'duration': 28.385, 'max_score': 875.066, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls875066.jpg'}], 'start': 385.307, 'title': 'Tensorflow in deep learning', 'summary': 'Covers the concept of tensors in tensorflow, its ability to run computations on multiple cpus and gpus, and its flexible architecture. it also introduces object detection in tensorflow, detailing input data provision, model training, feature extraction, and final output generation.', 'chapters': [{'end': 467.397, 'start': 385.307, 'title': 'Tensorflow: understanding tensors and data flow', 'summary': "Explains the concept of tensors in tensorflow, their representation in deep learning, the framework's ability to run computations on multiple cpus and gpus, and its flexible architecture enabling deployment across various platforms.", 'duration': 82.09, 'highlights': ["The concept of tensors in TensorFlow, their representation in deep learning, and the framework's ability to run computations on multiple CPUs and GPUs.", 'The flexible architecture of TensorFlow enabling easy deployment of computation across various platforms, including CPUs, GPUs, and TPUs.', 'The computation in TensorFlow is approached as a data flow graph, and computations are expressed as stateful data flow graphs.', 'Tensors are the standard way of representing data in deep learning, and they are simple arrays of numbers or functions that transform according to certain rules under a change of coordinates.']}, {'end': 896.979, 'start': 467.797, 'title': 'Tensorflow object detection', 'summary': 'Introduces object detection in tensorflow, detailing the process of input data provision, model training, feature extraction, and final output generation, with an emphasis on usage of tensorflow, tf record, protobuf, and coco api.', 'duration': 429.182, 'highlights': ['The TensorFlow object detection process involves input data provision, model training, feature extraction, and final output generation, where image conversion into a numpy array and usage of TF record is emphasized.', 'The process also includes the installation and usage of TensorFlow, protobuf, and Coco API, with specific instructions for setting up, compiling, and selecting a suitable model based on system resources and accuracy requirements.', 'The TensorFlow object detection model uses protobuf to configure model and training parameters, with specific instructions for downloading, extracting, and compiling protobuf libraries for usage.', 'The chapter provides an overview of the Coco dataset, including its image count, labeled images, object instances, categories, and other key statistics, highlighting its relevance in object detection processes.', 'The chapter emphasizes the significance of selecting a suitable model based on system resources and accuracy requirements, providing insights into the available models and their suitability based on GPU and RAM capabilities.']}], 'duration': 511.672, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls385307.jpg', 'highlights': ['The flexible architecture of TensorFlow enabling easy deployment of computation across various platforms, including CPUs, GPUs, and TPUs.', "The concept of tensors in TensorFlow, their representation in deep learning, and the framework's ability to run computations on multiple CPUs and GPUs.", 'The TensorFlow object detection process involves input data provision, model training, feature extraction, and final output generation, where image conversion into a numpy array and usage of TF record is emphasized.', 'The computation in TensorFlow is approached as a data flow graph, and computations are expressed as stateful data flow graphs.', 'The chapter emphasizes the significance of selecting a suitable model based on system resources and accuracy requirements, providing insights into the available models and their suitability based on GPU and RAM capabilities.']}, {'end': 1182.994, 'segs': [{'end': 1038.784, 'src': 'embed', 'start': 922.344, 'weight': 0, 'content': [{'end': 926.467, 'text': 'We are operating the collections and the various imports which are needed.', 'start': 922.344, 'duration': 4.123}, {'end': 930.595, 'text': 'Then we need to append the path of the object detection folder.', 'start': 926.874, 'duration': 3.721}, {'end': 941.678, 'text': 'And finally if the version of tensorflow is less than 1.4, we need to upgrade as the latest tutorial suppose the tensorflow object 1.4 and above.', 'start': 930.995, 'duration': 10.683}, {'end': 945.079, 'text': "So let's run this block by block.", 'start': 943.118, 'duration': 1.961}, {'end': 948.259, 'text': "So first of all, let's load all the libraries.", 'start': 945.799, 'duration': 2.46}, {'end': 953.901, 'text': 'Now next what we are going to do is import the object detection module some of the labels.', 'start': 949.56, 'duration': 4.341}, {'end': 961.404, 'text': 'which are the label map util which will be later used to provide the labels to the input images and based on that our model will be created.', 'start': 954.457, 'duration': 6.947}, {'end': 966.669, 'text': 'Now next what we are going to do is we are going to select which model to download.', 'start': 962.244, 'duration': 4.425}, {'end': 972.014, 'text': 'So for example here, we are using the SSD mobile net version 1 Coco 2017.', 'start': 967.109, 'duration': 4.905}, {'end': 977.619, 'text': 'So if you go back to the list of the models, you can select any of the given models here.', 'start': 972.014, 'duration': 5.605}, {'end': 982.835, 'text': 'but make sure your system should support the required amount of RAM.', 'start': 978.174, 'duration': 4.661}, {'end': 986.716, 'text': 'I should have the requirement of GPU to support the models which you are selecting.', 'start': 982.855, 'duration': 3.861}, {'end': 992.858, 'text': "So for this tutorial, I'm using a model which will give me the results faster.", 'start': 987.636, 'duration': 5.222}, {'end': 999.359, 'text': 'So all you need to do is provide the model name the model file and the download base from where it should download.', 'start': 993.338, 'duration': 6.021}, {'end': 1006.561, 'text': 'Now as I mentioned earlier that tensorflow works on the graph principle, which is the data flow graph.', 'start': 1000.78, 'duration': 5.781}, {'end': 1015.67, 'text': 'So what we are going to do is give the part to the detection graph which we are going to use here, which will be supported by this model,', 'start': 1007.324, 'duration': 8.346}, {'end': 1017.672, 'text': 'and then we are going to give the part to the labels.', 'start': 1015.67, 'duration': 2.002}, {'end': 1020.534, 'text': 'Now to download the models.', 'start': 1019.253, 'duration': 1.281}, {'end': 1029.181, 'text': 'We have this code which will take the URL and which will download this file and produce the frozen inference graph of that model,', 'start': 1020.614, 'duration': 8.567}, {'end': 1031.403, 'text': 'which is the SSD Coco mobile net.', 'start': 1029.181, 'duration': 2.222}, {'end': 1038.784, 'text': 'So once this has been done, We are going to load the graph which is the frozen inference graph into the memory.', 'start': 1033.223, 'duration': 5.561}], 'summary': 'Operating collections and imports, appending path, upgrading tensorflow, loading libraries, selecting and downloading model, downloading and loading frozen inference graph.', 'duration': 116.44, 'max_score': 922.344, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls922344.jpg'}, {'end': 1133.691, 'src': 'embed', 'start': 1107.611, 'weight': 3, 'content': [{'end': 1115.417, 'text': 'First of all, it detects all the boxes the detection mask and provides a certain box on the object it detects.', 'start': 1107.611, 'duration': 7.806}, {'end': 1125.084, 'text': 'And finally, we have the for loop, the mean for loop, in which will take the images from the test image path and open them,', 'start': 1116.818, 'duration': 8.266}, {'end': 1130.408, 'text': 'and one by one will take all the images and do the inference for a single image, one by one.', 'start': 1125.084, 'duration': 5.324}, {'end': 1133.691, 'text': 'So as you can see we are using the load image into numpy array.', 'start': 1130.829, 'duration': 2.862}], 'summary': 'The process detects boxes using a detection mask and runs inference on test images using a for loop.', 'duration': 26.08, 'max_score': 1107.611, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls1107611.jpg'}], 'start': 897.987, 'title': 'Tensorflow object detection', 'summary': 'Covers the setup of tensorflow object detection, including importing libraries, selecting a model, and ensuring compatibility. it also explains using the ssd coco mobilenet model for object detection, including downloading, loading into memory, converting images to numpy array, running inference for single and multiple images, and displaying the detected objects using matplotlib.', 'chapters': [{'end': 992.858, 'start': 897.987, 'title': 'Tensorflow object detection setup', 'summary': 'Outlines the process of setting up tensorflow object detection, including importing libraries, selecting a model, and ensuring tensorflow version compatibility.', 'duration': 94.871, 'highlights': ['The tutorial explains the process of importing necessary libraries such as numpy, OS, and TensorFlow, and updating TensorFlow to version 1.4 or above if necessary.', 'It details the import of the object detection module, label map utility, and the selection of a specific model (SSD MobileNet version 1 Coco 2017) for faster results.', "It emphasizes the importance of selecting a model compatible with the system's RAM and GPU requirements for efficient performance."]}, {'end': 1182.994, 'start': 993.338, 'title': 'Tensorflow object detection with ssd coco mobilenet', 'summary': 'Explains the process of setting up and using the ssd coco mobilenet model for object detection in tensorflow, including downloading, loading into memory, converting images to numpy array, running inference for single and multiple images, and displaying the detected objects using matplotlib.', 'duration': 189.656, 'highlights': ['Setting up and using the SSD Coco Mobilenet model for object detection in TensorFlow, including downloading, loading into memory, and converting images to numpy array.', 'The process of running inference for single and multiple images, including detecting boxes, detection mask, and providing detection score.', 'Explaining the use of tensorflow working on the graph principle, data flow graph, loading the graph into the memory, and using tf.graph method and tf.graph def to define the graph.']}], 'duration': 285.007, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls897987.jpg', 'highlights': ['The tutorial explains the process of importing necessary libraries such as numpy, OS, and TensorFlow, and updating TensorFlow to version 1.4 or above if necessary.', 'It details the import of the object detection module, label map utility, and the selection of a specific model (SSD MobileNet version 1 Coco 2017) for faster results.', 'Setting up and using the SSD Coco Mobilenet model for object detection in TensorFlow, including downloading, loading into memory, and converting images to numpy array.', 'The process of running inference for single and multiple images, including detecting boxes, detection mask, and providing detection score.', "It emphasizes the importance of selecting a model compatible with the system's RAM and GPU requirements for efficient performance.", 'Explaining the use of tensorflow working on the graph principle, data flow graph, loading the graph into the memory, and using tf.graph method and tf.graph def to define the graph.']}, {'end': 1633.586, 'segs': [{'end': 1238.707, 'src': 'embed', 'start': 1208.885, 'weight': 0, 'content': [{'end': 1212.467, 'text': 'it can detect objects in such a heavy background.', 'start': 1208.885, 'duration': 3.582}, {'end': 1218.771, 'text': 'as you can see, the person are so much camouflage in the background, but still it has managed to score that person.', 'start': 1212.467, 'duration': 6.304}, {'end': 1224.976, 'text': 'As you see here, it has detected an airplane person kite.', 'start': 1220.632, 'duration': 4.344}, {'end': 1229.139, 'text': 'Now you can use your own images.', 'start': 1226.897, 'duration': 2.242}, {'end': 1238.707, 'text': 'All you need to do is copy that images into the test image folder and use the naming convention provided here as the image one image two.', 'start': 1229.299, 'duration': 9.408}], 'summary': 'Object detection model successfully detects person, airplane, and kite in challenging backgrounds.', 'duration': 29.822, 'max_score': 1208.885, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls1208885.jpg'}, {'end': 1361.235, 'src': 'embed', 'start': 1331.661, 'weight': 2, 'content': [{'end': 1341.182, 'text': 'So what it will do is it will take the input of the video which is coming through the webcam and then We need to remove these two lines as well,', 'start': 1331.661, 'duration': 9.521}, {'end': 1349.368, 'text': "which is the PLT dot figure and the PLT dot image show, because we are not using matplotlib anymore and we'll be using the open CV.", 'start': 1341.182, 'duration': 8.186}, {'end': 1361.235, 'text': "So we'll do cv2.mshow and this is the name of the window which will come up and for instance, we are resizing the image to 800 is to 600 now.", 'start': 1349.588, 'duration': 11.647}], 'summary': 'Using opencv to process webcam video input, resizing image to 800x600.', 'duration': 29.574, 'max_score': 1331.661, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls1331661.jpg'}, {'end': 1452.073, 'src': 'embed', 'start': 1426.636, 'weight': 1, 'content': [{'end': 1433.381, 'text': 'And finally we need to run our final code block, which is the object detection which will call all the definitions which have provided earlier.', 'start': 1426.636, 'duration': 6.745}, {'end': 1445.029, 'text': 'So, guys, as you can see here, it has opened up a window of 800, cross, 600 and, as you can see, it has successfully detected the TV as 92% of TV,', 'start': 1434.922, 'duration': 10.107}, {'end': 1452.073, 'text': 'and keyword also has been detected at 90%.', 'start': 1445.029, 'duration': 7.044}], 'summary': 'Object detection code successfully detects tv at 92% and keyword at 90%.', 'duration': 25.437, 'max_score': 1426.636, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls1426636.jpg'}, {'end': 1519.15, 'src': 'embed', 'start': 1487.341, 'weight': 3, 'content': [{'end': 1493.927, 'text': 'but if you have a higher GPU system and the CPU is good, you can use more advanced model with the higher refresh rate.', 'start': 1487.341, 'duration': 6.586}, {'end': 1500.452, 'text': 'So as you can see here, it has detected clock the watch as clock.', 'start': 1494.607, 'duration': 5.845}, {'end': 1507.877, 'text': 'And the moment it receives any part of human body in the frame as you can see.', 'start': 1502.093, 'duration': 5.784}, {'end': 1512.561, 'text': 'If my hand is there it will detect it as a person.', 'start': 1509.879, 'duration': 2.682}, {'end': 1519.15, 'text': 'Okay You can see it is detecting the TV the cup.', 'start': 1514.763, 'duration': 4.387}], 'summary': 'Higher gpu system allows for advanced model with higher refresh rate, detecting various objects and body parts.', 'duration': 31.809, 'max_score': 1487.341, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls1487341.jpg'}], 'start': 1183.054, 'title': 'Object detection methods', 'summary': 'Covers object detection using pre-trained models achieving 94% accuracy, live object detection using opencv, and object detection with tensorflow achieving 92% accuracy in detecting tv and 90% accuracy in detecting keywords.', 'chapters': [{'end': 1425.627, 'start': 1183.054, 'title': 'Object detection and live detection using opencv', 'summary': 'Covers object detection using a pre-trained model which achieved a 94% detection score and live object detection using opencv, with explanations on model selection, code modifications, and necessary libraries.', 'duration': 242.573, 'highlights': ['Object detection achieved a 94% detection score, identifying a dog, person, kite, and airplane in various scenarios, even in heavy background camouflage.', "Live object detection using OpenCV requires code modifications such as using 'cv2.mshow' instead of 'PLT.figure' and 'PLT.image show', and handling input from the webcam.", 'Model selection for live object detection involves choosing a model that balances speed and accuracy based on GPU capability and image refresh rate.', 'Object detection allows users to add custom images by copying them to the test image folder and using the provided naming convention.']}, {'end': 1633.586, 'start': 1426.636, 'title': 'Tensorflow object detection', 'summary': 'Demonstrates the object detection using tensorflow, achieving 92% accuracy in detecting tv and 90% accuracy in detecting keywords, while also discussing the potential for using more advanced models with higher refresh rates.', 'duration': 206.95, 'highlights': ['The object detection achieved 92% accuracy in detecting the TV and 90% accuracy in detecting keywords.', 'The tutorial provides insights into using Tensorflow models generated with different datasets, such as coco and kitty, for real-time object detection and image input.', "Discussion on the potential for using more advanced models with higher refresh rates, based on the system's GPU and CPU capabilities.", "Upcoming video topics include creating custom image classifiers using one's own image dataset, splitting the dataset for training and testing, and creating a personalized model."]}], 'duration': 450.532, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/wh7_etX91ls/pics/wh7_etX91ls1183054.jpg', 'highlights': ['Object detection achieved a 94% detection score, identifying a dog, person, kite, and airplane in various scenarios, even in heavy background camouflage.', 'The object detection achieved 92% accuracy in detecting the TV and 90% accuracy in detecting keywords.', "Live object detection using OpenCV requires code modifications such as using 'cv2.mshow' instead of 'PLT.figure' and 'PLT.image show', and handling input from the webcam.", 'Model selection for live object detection involves choosing a model that balances speed and accuracy based on GPU capability and image refresh rate.']}], 'highlights': ['Object detection achieved a 94% detection score, identifying a dog, person, kite, and airplane in various scenarios, even in heavy background camouflage.', 'Object detection achieved 92% accuracy in detecting the TV and 90% accuracy in detecting keywords.', 'Recent advancements in deep learning have significantly improved the performance of object detection, enabling real-time use cases.', 'Object detection allows for the recognition, detection, and localization of multiple objects within an image, providing a better understanding of the image as a whole.', 'Object detection is used in various industrial applications such as facial recognition for people counting, industrial processes, inventory management, and self-driving cars, emphasizing its role in security and banking industries.']}