CNN

A real-time human counting system with age and gender classification.

Real-time Gender and Age Recognition from Audio

En este trabajo se varían distintos hiperparámetros de una CNN para mejorar la clasificación multiclase de imágenes

A project focused on weather classification using advanced deep learning techniques, specifically leveraging TensorFlow and a custom Convolutional Neural Network (CNN). The project involved the integration of four diverse weather datasets, namely ACDC, MWD, UAVid, and Syndrone, covering various weather conditions, including clear sky, cloudy, rainy, and sunny weather. Developed a custom CNN architecture using TensorFlow's Keras API, incorporating convolutional layers for feature extraction and dense layers for classification.

Prediction of age from X-Ray images of hand bones using deep learning models. We used 3 models: Shallow, ResNet50, and InceptionV4. The best result achieved was with a mean absolute error of 10 months using InceptionV4. The preprocessing of data included computer vision techniques like CLAHE filter and reducing channels, and also creativities such as using the Google MediaPipe library to detect hands and crop on them.

Introduction to AI & Data Analysis: Classification of Iris-dataset with classic Perceptron Classification of MNIST-dataset with MLP and CNN

MLP and CNN to classify TinyImage30 dataset. Fine Tuned models. Applied Grad-Cam to identify parts of the image that highly impact the classification based on model convolution gradients. Feature-2-Seq RNN encoder/decoder network trained on the COCO dataset. The produced model is able to predict reasonable captions for provided test images.

Deep learning (CNN) deployment pipeline

Develop a cool feature in the smart-TV that can recognize five different gestures performed by the user which will help users control the TV without using a remote.

To build a CNN based model which can accurately detect melanoma. Melanoma is a type of cancer that can be deadly if not detected early. It accounts for 75% of skin cancer deaths. A solution which can evaluate images and alert the dermatologists.

This project deals with human detection from IR-UWB radar sensors. In this project, we implement a custom Convolutional Neural Network with TensorFlow to classify reflected signals from people queuing and walking in a closed space.

TensorFlow examples for Artificial Neural Networks course including: MLP with softmax output layer; MLP and CNN for MNIST dataset; CNN for CIFAR-10 dataset with data augmentation; LSTM with CNN layer for IMDB sentiment classification task.

Creating an Image Search Engine, which uses a vanilla CNN to extract features from an image, runs them through an Autoencoder for extracting relevant features, clusters them using a K-Means clustering, then gets similar images using Siamese networks.

Python code for detection and recognition of 13 VIPs. Project for Digital Forensics A.A 19/20 course for UNIPD.

Time series classification to identify relapse / non-relapse after cessation of leukemia treatment

Experiments for the paper: "Improving Machine Hearing on Limited Data Sets"

Comparison of standard Gabor transform with Gabor scattering audio signal representations used as an input of convolutional neural network on multiclass classification task. Experiments conducted on synthetic and GoodSounds data.

Various approaches for the Data Science Bowl 2018 Kaggle Competition. https://datasciencebowl.com : Segmenting nuclei to speed up tedious tasks in biomedical engineering efforts to find new cures.