Documenting my study of" An Introduction to Statistical Learning with Python " book - theory, code, exercises, notes and my progress all the way through

using Neural Networks (SSD) on Tensorflow. This repo documents steps and scripts used to train a hand detector using Tensorflow (Object Detection API). As with any DNN based task, the most expensive (and riskiest) part of the process has to do with finding or creating the right (annotated) dataset. I was interested mainly in detecting hands on a table (egocentric view point). I experimented first with the [Oxford Hands Dataset](http://www.robots.ox.ac.uk/~vgg/data/hands/) (the results were not good). I then tried the [Egohands Dataset](http://vision.soic.indiana.edu/projects/egohands/) which was a much better fit to my requirements. The goal of this repo/post is to demonstrate how neural networks can be applied to the (hard) problem of tracking hands (egocentric and other views). Better still, provide code that can be adapted to other uses cases. If you use this tutorial or models in your research or project, please cite [this](#citing-this-tutorial). Here is the detector in action. <img src="images/hand1.gif" width="33.3%"><img src="images/hand2.gif" width="33.3%"><img src="images/hand3.gif" width="33.3%"> Realtime detection on video stream from a webcam . <img src="images/chess1.gif" width="33.3%"><img src="images/chess2.gif" width="33.3%"><img src="images/chess3.gif" width="33.3%"> Detection on a Youtube video. Both examples above were run on a macbook pro **CPU** (i7, 2.5GHz, 16GB). Some fps numbers are: | FPS | Image Size | Device| Comments| | ------------- | ------------- | ------------- | ------------- | | 21 | 320 * 240 | Macbook pro (i7, 2.5GHz, 16GB) | Run without visualizing results| | 16 | 320 * 240 | Macbook pro (i7, 2.5GHz, 16GB) | Run while visualizing results (image above) | | 11 | 640 * 480 | Macbook pro (i7, 2.5GHz, 16GB) | Run while visualizing results (image above) | > Note: The code in this repo is written and tested with Tensorflow `1.4.0-rc0`. Using a different version may result in [some errors](https://github.com/tensorflow/models/issues/1581). You may need to [generate your own frozen model](https://pythonprogramming.net/testing-custom-object-detector-tensorflow-object-detection-api-tutorial/?completed=/training-custom-objects-tensorflow-object-detection-api-tutorial/) graph using the [model checkpoints](model-checkpoint) in the repo to fit your TF version. **Content of this document** - Motivation - Why Track/Detect hands with Neural Networks - Data preparation and network training in Tensorflow (Dataset, Import, Training) - Training the hand detection Model - Using the Detector to Detect/Track hands - Thoughts on Optimizations. > P.S if you are using or have used the models provided here, feel free to reach out on twitter ([@vykthur](https://twitter.com/vykthur)) and share your work! ## Motivation - Why Track/Detect hands with Neural Networks? There are several existing approaches to tracking hands in the computer vision domain. Incidentally, many of these approaches are rule based (e.g extracting background based on texture and boundary features, distinguishing between hands and background using color histograms and HOG classifiers,) making them not very robust. For example, these algorithms might get confused if the background is unusual or in situations where sharp changes in lighting conditions cause sharp changes in skin color or the tracked object becomes occluded.(see [here for a review](https://www.cse.unr.edu/~bebis/handposerev.pdf) paper on hand pose estimation from the HCI perspective) With sufficiently large datasets, neural networks provide opportunity to train models that perform well and address challenges of existing object tracking/detection algorithms - varied/poor lighting, noisy environments, diverse viewpoints and even occlusion. The main drawbacks to usage for real-time tracking/detection is that they can be complex, are relatively slow compared to tracking-only algorithms and it can be quite expensive to assemble a good dataset. But things are changing with advances in fast neural networks. Furthermore, this entire area of work has been made more approachable by deep learning frameworks (such as the tensorflow object detection api) that simplify the process of training a model for custom object detection. More importantly, the advent of fast neural network models like ssd, faster r-cnn, rfcn (see [here](https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md#coco-trained-models-coco-models) ) etc make neural networks an attractive candidate for real-time detection (and tracking) applications. Hopefully, this repo demonstrates this. > If you are not interested in the process of training the detector, you can skip straight to applying the [pretrained model I provide in detecting hands](#detecting-hands). Training a model is a multi-stage process (assembling dataset, cleaning, splitting into training/test partitions and generating an inference graph). While I lightly touch on the details of these parts, there are a few other tutorials cover training a custom object detector using the tensorflow object detection api in more detail[ see [here](https://pythonprogramming.net/training-custom-objects-tensorflow-object-detection-api-tutorial/) and [here](https://towardsdatascience.com/how-to-train-your-own-object-detector-with-tensorflows-object-detector-api-bec72ecfe1d9) ]. I recommend you walk through those if interested in training a custom object detector from scratch. ## Data preparation and network training in Tensorflow (Dataset, Import, Training) **The Egohands Dataset** The hand detector model is built using data from the [Egohands Dataset](http://vision.soic.indiana.edu/projects/egohands/) dataset. This dataset works well for several reasons. It contains high quality, pixel level annotations (>15000 ground truth labels) where hands are located across 4800 images. All images are captured from an egocentric view (Google glass) across 48 different environments (indoor, outdoor) and activities (playing cards, chess, jenga, solving puzzles etc). <img src="images/egohandstrain.jpg" width="100%"> If you will be using the Egohands dataset, you can cite them as follows: > Bambach, Sven, et al. "Lending a hand: Detecting hands and recognizing activities in complex egocentric interactions." Proceedings of the IEEE International Conference on Computer Vision. 2015. The Egohands dataset (zip file with labelled data) contains 48 folders of locations where video data was collected (100 images per folder). ``` -- LOCATION_X -- frame_1.jpg -- frame_2.jpg ... -- frame_100.jpg -- polygons.mat // contains annotations for all 100 images in current folder -- LOCATION_Y -- frame_1.jpg -- frame_2.jpg ... -- frame_100.jpg -- polygons.mat // contains annotations for all 100 images in current folder ``` **Converting data to Tensorflow Format** Some initial work needs to be done to the Egohands dataset to transform it into the format (`tfrecord`) which Tensorflow needs to train a model. This repo contains `egohands_dataset_clean.py` a script that will help you generate these csv files. - Downloads the egohands datasets - Renames all files to include their directory names to ensure each filename is unique - Splits the dataset into train (80%), test (10%) and eval (10%) folders. - Reads in `polygons.mat` for each folder, generates bounding boxes and visualizes them to ensure correctness (see image above). - Once the script is done running, you should have an images folder containing three folders - train, test and eval. Each of these folders should also contain a csv label document each - `train_labels.csv`, `test_labels.csv` that can be used to generate `tfrecords` Note: While the egohands dataset provides four separate labels for hands (own left, own right, other left, and other right), for my purpose, I am only interested in the general `hand` class and label all training data as `hand`. You can modify the data prep script to generate `tfrecords` that support 4 labels. Next: convert your dataset + csv files to tfrecords. A helpful guide on this can be found [here](https://pythonprogramming.net/creating-tfrecord-files-tensorflow-object-detection-api-tutorial/).For each folder, you should be able to generate `train.record`, `test.record` required in the training process. ## Training the hand detection Model Now that the dataset has been assembled (and your tfrecords), the next task is to train a model based on this. With neural networks, it is possible to use a process called [transfer learning](https://www.tensorflow.org/tutorials/image_retraining) to shorten the amount of time needed to train the entire model. This means we can take an existing model (that has been trained well on a related domain (here image classification) and retrain its final layer(s) to detect hands for us. Sweet!. Given that neural networks sometimes have thousands or millions of parameters that can take weeks or months to train, transfer learning helps shorten training time to possibly hours. Tensorflow does offer a few models (in the tensorflow [model zoo](https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md#coco-trained-models-coco-models)) and I chose to use the `ssd_mobilenet_v1_coco` model as my start point given it is currently (one of) the fastest models (read the SSD research [paper here](https://arxiv.org/pdf/1512.02325.pdf)). The training process can be done locally on your CPU machine which may take a while or better on a (cloud) GPU machine (which is what I did). For reference, training on my macbook pro (tensorflow compiled from source to take advantage of the mac's cpu architecture) the maximum speed I got was 5 seconds per step as opposed to the ~0.5 seconds per step I got with a GPU. For reference it would take about 12 days to run 200k steps on my mac (i7, 2.5GHz, 16GB) compared to ~5hrs on a GPU. > **Training on your own images**: Please use the [guide provided by Harrison from pythonprogramming](https://pythonprogramming.net/training-custom-objects-tensorflow-object-detection-api-tutorial/) on how to generate tfrecords given your label csv files and your images. The guide also covers how to start the training process if training locally. [see [here] (https://pythonprogramming.net/training-custom-objects-tensorflow-object-detection-api-tutorial/)]. If training in the cloud using a service like GCP, see the [guide here](https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/running_on_cloud.md). As the training process progresses, the expectation is that total loss (errors) gets reduced to its possible minimum (about a value of 1 or thereabout). By observing the tensorboard graphs for total loss(see image below), it should be possible to get an idea of when the training process is complete (total loss does not decrease with further iterations/steps). I ran my training job for 200k steps (took about 5 hours) and stopped at a total Loss (errors) value of 2.575.(In retrospect, I could have stopped the training at about 50k steps and gotten a similar total loss value). With tensorflow, you can also run an evaluation concurrently that assesses your model to see how well it performs on the test data. A commonly used metric for performance is mean average precision (mAP) which is single number used to summarize the area under the precision-recall curve. mAP is a measure of how well the model generates a bounding box that has at least a 50% overlap with the ground truth bounding box in our test dataset. For the hand detector trained here, the mAP value was **0.9686@0.5IOU**. mAP values range from 0-1, the higher the better. <img src="images/accuracy.jpg" width="100%"> Once training is completed, the trained inference graph (`frozen_inference_graph.pb`) is then exported (see the earlier referenced guides for how to do this) and saved in the `hand_inference_graph` folder. Now its time to do some interesting detection. ## Using the Detector to Detect/Track hands If you have not done this yet, please following the guide on installing [Tensorflow and the Tensorflow object detection api](https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/installation.md). This will walk you through setting up the tensorflow framework, cloning the tensorflow github repo and a guide on - Load the `frozen_inference_graph.pb` trained on the hands dataset as well as the corresponding label map. In this repo, this is done in the `utils/detector_utils.py` script by the `load_inference_graph` method. ```python detection_graph = tf.Graph() with detection_graph.as_default(): od_graph_def = tf.GraphDef() with tf.gfile.GFile(PATH_TO_CKPT, 'rb') as fid: serialized_graph = fid.read() od_graph_def.ParseFromString(serialized_graph) tf.import_graph_def(od_graph_def, name='') sess = tf.Session(graph=detection_graph) print("> ====== Hand Inference graph loaded.") ``` - Detect hands. In this repo, this is done in the `utils/detector_utils.py` script by the `detect_objects` method. ```python (boxes, scores, classes, num) = sess.run( [detection_boxes, detection_scores, detection_classes, num_detections], feed_dict={image_tensor: image_np_expanded}) ``` - Visualize detected bounding detection_boxes. In this repo, this is done in the `utils/detector_utils.py` script by the `draw_box_on_image` method. This repo contains two scripts that tie all these steps together. - detect_multi_threaded.py : A threaded implementation for reading camera video input detection and detecting. Takes a set of command line flags to set parameters such as `--display` (visualize detections), image parameters `--width` and `--height`, videe `--source` (0 for camera) etc. - detect_single_threaded.py : Same as above, but single threaded. This script works for video files by setting the video source parameter videe `--source` (path to a video file). ```cmd # load and run detection on video at path "videos/chess.mov" python detect_single_threaded.py --source videos/chess.mov ``` > Update: If you do have errors loading the frozen inference graph in this repo, feel free to generate a new graph that fits your TF version from the model-checkpoint in this repo. Use the [export_inference_graph.py](https://github.com/tensorflow/models/blob/master/research/object_detection/export_inference_graph.py) script provided in the tensorflow object detection api repo. More guidance on this [here](https://pythonprogramming.net/testing-custom-object-detector-tensorflow-object-detection-api-tutorial/?completed=/training-custom-objects-tensorflow-object-detection-api-tutorial/). ## Thoughts on Optimization. A few things that led to noticeable performance increases. - Threading: Turns out that reading images from a webcam is a heavy I/O event and if run on the main application thread can slow down the program. I implemented some good ideas from [Adrian Rosebuck](https://www.pyimagesearch.com/2017/02/06/faster-video-file-fps-with-cv2-videocapture-and-opencv/) on parrallelizing image capture across multiple worker threads. This mostly led to an FPS increase of about 5 points. - For those new to Opencv, images from the `cv2.read()` method return images in [BGR format](https://www.learnopencv.com/why-does-opencv-use-bgr-color-format/). Ensure you convert to RGB before detection (accuracy will be much reduced if you dont). ```python cv2.cvtColor(image_np, cv2.COLOR_BGR2RGB) ``` - Keeping your input image small will increase fps without any significant accuracy drop.(I used about 320 x 240 compared to the 1280 x 720 which my webcam provides). - Model Quantization. Moving from the current 32 bit to 8 bit can achieve up to 4x reduction in memory required to load and store models. One way to further speed up this model is to explore the use of [8-bit fixed point quantization](https://heartbeat.fritz.ai/8-bit-quantization-and-tensorflow-lite-speeding-up-mobile-inference-with-low-precision-a882dfcafbbd). Performance can also be increased by a clever combination of tracking algorithms with the already decent detection and this is something I am still experimenting with. Have ideas for optimizing better, please share! <img src="images/general.jpg" width="100%"> Note: The detector does reflect some limitations associated with the training set. This includes non-egocentric viewpoints, very noisy backgrounds (e.g in a sea of hands) and sometimes skin tone. There is opportunity to improve these with additional data. ## Integrating Multiple DNNs. One way to make things more interesting is to integrate our new knowledge of where "hands" are with other detectors trained to recognize other objects. Unfortunately, while our hand detector can in fact detect hands, it cannot detect other objects (a factor or how it is trained). To create a detector that classifies multiple different objects would mean a long involved process of assembling datasets for each class and a lengthy training process. > Given the above, a potential strategy is to explore structures that allow us **efficiently** interleave output form multiple pretrained models for various object classes and have them detect multiple objects on a single image. An example of this is with my primary use case where I am interested in understanding the position of objects on a table with respect to hands on same table. I am currently doing some work on a threaded application that loads multiple detectors and outputs bounding boxes on a single image. More on this soon.

This repository tracks my progress through Dr. Angela Yu's 100 Days of Code: The Complete Python Pro Bootcamp for 2023 course.

This Python folder is a structured learning workspace that documents my progress from beginner to more advanced programming concepts through daily practice projects.

1-Project Title “Library Management System (L_I_B_R_A)” The “Library Management System” has been developed to override the problems prevailing in the practicing manual system. This software is supported to eliminate and in some cases reduce the hardship faced by this existing system. Moreover this system is designed for the particular need of the institution to carry out operating in a smooth and effective manner. 2-Domain Library management institutional management non-profitable organization. 3-Problem of Statement In our existing system all the transaction of books are done manually, so taking more time for a transaction like borrowing a book or returning a book and also for searching of member and books. Another major disadvantage is that to preparing the list of book borrowed and the available book in the library will take more time, currently it is doing as a day process for verifying all records. So after conducting he feasibility study we decided to make the manual library management system to be computerised. Proposed system is an automated library management system. Through our software user can add member, edit information, borrow and return books in quick time. Some of the problems being faced in faced in manual system are as follows: • Fast report generation is not possible. • Tracing a book is difficult. • Information about issue/return of the books is not properly maintained. • No central database can be created as information is not available in database. All the manual difficulty in managing the Library have been rectified by implementing computerization. This application is reduced as much as possible to avoid errors while entering the data. It also provides error message while entering invalid data. No, formal knowledge is needed for the user to use this system. Thus by this all it is user-friendly. Library Management System as described above can lead error free, secure, reliable and fast management system. It can assist the user to concentrate on other activities rather to concentrate on the record keeping. Thus it will help organisation in better utilization of resources. So that’s why I can choose this topic to make it simple. is a sub-discipline of issues faced by libraries and library management professionals. Library management encompasses normal managerial tasks, as well as intellectual that focuses on specific freedom and fundraising responsibilities. Issues faced in library management frequently overlap with those faced in managing Title: L_I_B_R_A Page 5 of 16 TMU-CCSIT Version 1. 4-Project Description The software to be produced is on Library Management System. A library card will also be provided to the customers who visit daily. A person can also borrow the book for particular days. All the information will be entered in the system. If the person doesn’t return the due book. Admin has the authority to add, delete or modify the details of the book available to/from the system. He also has the authority to provide username and password for the receptionist. He can also add the details of the book purchased from shops along with the shop name. Project plan Requirement Design Process description implementation STATE DIAGRAM : LIBRARIAN OBJECT Title: L_I_B_R_A Page 6 of 16 TMU-CCSIT Version 1. 4.1-Scope of the Work This project is helpful to track all the book and library information and to rate the maximum number of books, the students wished to allot books. The software will be able to handle all the necessary information related to the library. From a librarian perspective, the Library Management System Project enhanced searchable database for the search books, managing library members, issuing and receiving books . • Search Books, Managing Library Members, Issuing and Receiving Books: An enhanced atomized system is developed to maintain Books, Authors, Issuing and Receiving Books and maintain the history of transaction. • To utilize resources in an efficient manner by increasing their productivity through automation. • It satisfies the user requirement. Title: L_I_B_R_A Page 7 of 16 TMU-CCSIT Version 1. 4.2-Project Modules • Books: This module consist the details of the books available in library and their categories. Title: L_I_B_R_A Page 8 of 16 TMU-CCSIT Version 1. • Member Account details To issue an book from the library, one should have a account in the library. The registration contains all the details about the member like registration number, name, address, contact number etc.. • Book Request: This module is used by the member to request a book from the library. The search can be performed by using name of the book, author name, and subject name. Title: L_I_B_R_A Page 9 of 16 TMU-CCSIT Version 1. • Issue of books: This module is used by the librarian to issue a book based on the request made by the members. • Returning Books: In this module the librarian maintains the details of the books returned by the member, which also includes the fine details, damage book details, lost book details. • History: In this module the member can view the details about the previous issued books, requested books and returned books etc. • Reports: This module includes the details about the issued books, returned books, member reports, fine reports, or any damage to the book or details of the book which are not returned. Title: L_I_B_R_A Page 10 of 16 TMU-CCSIT Version 1. 5-Implementation Methodology In this I am trying to give an Idea of “How I can implement the library management system” . FUNCTIONAL DECOMPOSITION OF LIBRARY MANAGEMENT SYSTEM CLASS DIAGRAM OF LIBRARY MANAGEMENT SYSTEM Title: L_I_B_R_A Page 11 of 16 TMU-CCSIT Version 1. DFD OF LIBRARY MANAGEMENT SYSTEM ER DIAGRAM OF LIBRARY MANAGEMENT SYSYTEM Title: L_I_B_R_A Page 12 of 16 TMU-CCSIT Version 1. DATABASE OF LIBRARY MANAGEMENT SYSTEM Title: L_I_B_R_A Page 13 of 16 TMU-CCSIT Version 1. 6-Technologies to be used 6.1 -Software Platform a) Front-end ----python (3.8) ----tk-inter (GUI) b) Back-end -----sqlite (database) 6.2 -Hardware Platform RAM — 8 GB Hard Disk — not used OS — Mac OS (Mojave-10.14.6) Editor — idle (Available with python package) Processor — 1.8 GHz intel core i5 6.3 -Tools No tool used. 7-Advantages of this Project Our proposed system has the following advantages. User friendly interface Fast access to database Less error More storage capacity Search facility Look and feel environment Quick transaction 8-Future Scope and further enhancement of the Project o In future we can make this application online so that members will be able to search the book from any places as well as can send book request. o Book reading facility can be provided through on-line. o In the area of data security and system security. o Provide more online tips and help. o Implementation of ISBN BAR code reader Title: L_I_B_R_A Page 14 of 16 TMU-CCSIT Version 1. 9-Project Repository Location S# Project Artifacts (softcopy) Location Verified by Project Guide Verified by Lab In-Charge 1. Project Synopsis Report (Final Version) https://s.docworkspace.com/d/AEsSC- 7eqLpR6Z6S_OSdFA Tushar mehrotra Name and Signature 2. Project Progress updates Name and Signature Name and Signature 3. Project Requirement specifications Name and Signature Name and Signature 4. Project Report (Final Version) Name and Signature Name and Signature 5. Test Repository Name and Signature Name and Signature 6. Any other document, give details Name and Signature Name and Signature 10-Team Details Project Name Course Name Student ID Student Name Role Signature LIBRARY MANAGEMENT SYSTEM INDUSTRIAL TRAINING(PYTHON) (ECS 509) TCA1809026 UTTKARSH PARMAR Developer 11-Conclusion “Library Management System” allows the user to store the book details and the customer details. This software package allows storing the details of all the data related to library. The system is strong enough to withstand regressive yearly operations under conditions where the Title: L_I_B_R_A Page 15 of 16 TMU-CCSIT Version 1. database is maintained and cleared over a certain time of span. The implementation of the system in the organization will considerably reduce data entry, time and also provide readily calculated reports. 12-References • Website http://www.wikipedia.com http://www.sololearn.com And also my mentor from Ducat (Noida) • Book Python-basic-handbook ( writer- vivek Krishnamoorthy, jay Parmar, mario pisa pena)

This Python folder is a structured learning workspace that documents my progress from beginner to more advanced programming concepts through daily practice projects.