A Fast Deep Learning Model to Upsample Low Resolution Videos to High Resolution at 30fps

Implemented 3 neural network architectures: 1) Combination of RNN LSTM nodes and CNN, 2) CNN with residual blocks similar to ResNet, 3) Deep RNN LSTM network; and compared their performance to detect 12 speech commands.

Weed Detection in Sugar Beet Plants

Graduation Project. Applying Generative Adversarial Networks(GAN) with Residual-In-Residual(RIR) blocks.

This is an implementation of FRED-Net using keras.

Python Keras CNN Implementations

Semantic segmentation for brain tumors

Fast bare-bones implementation of convolutional layers, residual blocks, Adam optimizer, backpropagation and custom accuracy and loss functions (F1 score per pixel and binary-crossentropy)

Attention Residual UNet for vein image segmentation in the field of biometric identification

Emotion and Facial Key-Point Detection Classify emotions and detect facial key-points using deep learning! This project combines CNNs and Residual Blocks to predict 15 facial key-points and categorize facial expressions into five emotions: Angry, Disgust, Sad, Happy, and Surprise.

Deep learning model to predict the normal flow between two consecutive frames, being the normal flow the projection of the optical flow on the gradient directions.