This python implemented project primarily makes use of OpenCV & Imutils libraries for object recognition with the aid of machine learning to detect humans in still digital images.
This project aims to tackle the task of human recognition in still images. The human body has similar body structure regardless of gender, build, race etc. At the basic level, we all have a head, arms, legs, hip etc. We can employ computer vision to extract these features and feed them to machine learning algorithms to detect and track humans in images.
I propose a human detection and tracking scheme based on the pre-trained Histogram of Oriented Gradients (HOG) Descriptor and Linear Support Vector Machine (SVM) included in the OpenCV library. This model is based on Dalal-Triggs algorithm, which automatically detects pedestrians in images, and can be used for detection in both images and video stream; however, the scope of my project is limited to images only.