Siddhant Thendral Arasu

https://githubcom/Siddhant-Thendral-Arasu/Siddhant-Thendral-Arasu/

CNN Sign Language Digit Classifier

Developed a CNN to classify American Sign Language digits (0–9) using the Sign Language MNIST dataset Applied data preprocessing, dropout regularization, and softmax output to improve prediction accuracy

Car Detection Using Sliding Window & YOLO Concepts

Implemented a two-part vehicle detection system: a basic sliding window classifier and a YOLO-inspired pipeline Used OpenCV and HOG features for candidate detection, then trained a CNN to classify image regions

Cat Image Classifier

Built and trained a convolutional neural network to distinguish cat images from non-cat images Performed model tuning and loss monitoring using TensorFlow/Keras

Chest X-Ray Pneumonia Classification

Trained a binary image classifier to detect pneumonia in chest X-rays Applied convolutional layers with batch normalization and max pooling, achieving high accuracy on medical image data

Road Scene Semantic Segmentation

Implemented a fully convolutional network for segmenting urban road scenes into classes like road, vehicle, pedestrian, and sky Adapted U-Net architecture and trained on Cityscapes-style dataset