Classification of Objects using Deep Learning Model

Training a Deep learning model to classify images of bricks, balls or cylinders against a cluttered background using the pre-trained Alexnet model in fastai/pytorch, to perform classification, detection and segmentation simultaneously.

Dataset

Dataset includes 1k images of bricks, balls and cylinders placed in with other surrounding objects.

Implementation and Results

Implemented methods like classification of images, creation of the bounding boxes around particular objects and segmentation of the images for easier analysis of different objects. Training the fastai model for multi-task learning using bbc–1k dataset which includes images of bricks, balls and cylinders. Model achieved 95% accuracy.

Technologies

numpy, pandas, openCV, scikit-learn, scipy, python, pytorch, fastai, google colab