Deep learning (also known as deep structured learning) is part of machine learning methods based on artificial neural networks with representation learning. Learning can be supervised, semi-supervised or unsupervised
The ability to recognise a person face by looking at the digital image.
The ability to detect and locate human face.
Social distancing implies that people should physically distance themselves from one another.
Crossing a certain threshold(6 feet) can cause a violation.
- Deep Learning out perform other techniques if the data size is large. But with small data size, traditional Machine Learning algorithms are preferable.
- Deep Learning techniques need to have high end infrastructure to train in reasonable time.
- When there is lack of domain understanding for feature introspection, Deep Learning techniques outshines others as you have to worry less about feature engineering.
- Deep Learning really shines when it comes to complex problems such as image classification, natural language processing, and speech recognition.
- To do object detection, I used the Yolov6 model.
Object Detection can be performed on:
* Image.
* Video.
* WebCam.
## Topics Covered:
1. Automatically detect lane markings in images. <br />
2. Detect cars and pedestrians using a trained classifier and with SVM. <br />
3. Classify traffic signs using Convolutional Neural Networks. <br />
4. Identify other vehicles in images using template matching.<br />
5. Build deep neural networks with Tensorflow and Keras.<br />
6. Analyze and visualize data with Numpy, Pandas, Matplotlib, and Seaborn.<br />
7. Process image data using OpenCV.<br />
8. Calibrate cameras in Python, correcting for distortion.<br />
9. Sharpen and blur images with convolution.<br />
10. Detect edges in images with Sobel, Laplace, and Canny.<br />
11. Transform images through translation, rotation, resizing, and perspective transform.<br />
12. Extract image features with HOG.<br />
13. Detect object corners with Harris.<br />
14. Classify data with machine learning techniques including regression, decision trees, Naive Bayes, and SVM.<br />
15. Classify data with artificial neural networks and deep learning.<br />
Extract pixel from an image. For eg: If we want to extract white color. Then extract the pixels that are close to 255.
Steps:
- Import all the necessary libraries.
- Read , Convert and display image.
- Extract the required pixel from image.
Check out the result here.
With TensorFlow.js we can develop ML models in JavaScript, and use ML directly in the browser or in Node.js.
- Run existing models.
- Retrain existing models.
- Develop ML with JS.
- Simple Face Detection: Detects faces in. images using a 'Single Shot Detector' Architecture with a custom encoder (BlazeFace).
Most of the time, I just experimented with object detection and recognition, which creates a bounding box around specific detected objects in an image. However, I later came across another technique that can give a precise outline of an object that has been detected in an image. This technique is known as Image Segmentation.
The automatic skin disorders classification can help people in identifying skin disorders that occur and immediately consult with medical personnel to get appropriate medical treatment.
Object identification is by far the most important use of computer vision.
However, identifying small objects and inference on large images remain significant challenges.
The task of recognizing and localizing objects that are small in size inside digital images is referred to as small object detection.
I've used the SAHI library in conjunction with the YOLOv8 model to produce outstanding enhancements in object recognition results.
SAHI is an acronym that stands for Slicing Aided Hyper Inference.
SAHI uses the power of slicing-aided inference and fine-tuning approaches to revolutionize object detection.
In this section, I've compared small object detection results obtained without and with SAHI.
SAHI delivers a game-changing slicing-assisted fine-tuning approach that raises detection accuracy to unprecedented heights.
![Screenshot 2023-09-09 at 8 27 39 PM](https://private-user-images.githubusercontent.com/58945964/266807937-cad70356-3256-402e-a1fe-199f41f1c570.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.e2zCqfcclhf6hC27YE5Kd7kcz73K-OogbhxBKxonNgQ)