| ID | Project | Description |
|---|---|---|
| 0001 | CNN From Scratch Tenserflow | Building a Simple CNN Model in Tensorflow (First time experimenting CNNs in Tensorflow). |
| 0002 | CNN Fashion Image | Building a CNN Model for the Fashion-MNIST Dataset. |
| 0003 | CNN Hand Gesture | Building a Hand Gesture Recognition for American Sign Language. |
| 0004 | Facial Expression Recognition | Facial Expression Recognition on the FER Dataset. |
| 0005 | Image Colorization | Image Colorization from black and white photos to colored photos. |
| 0006 | Audio Classification | Audio Classfication with the Gtzan Dataset. |
| 0007 | Facial Key Points Detection | Building a system to identify key facial landmarks. |
| 0008 | OCR Recognition | Optical Character Recognition using Pytorch |
| 0009 | Scene Classification | System built where scenes from photographs are categorically classified. |
| 0010 | Traffic Sign Recoginition | Traffic Sign Recognition. |
| 0011 | Real Time Shape Contour Detection | Detect contours of shapes from a live video feed. |
| 0012 | Real Time Color Tracker | Detect and Track color from the HSV scale from a live video feed. |
| 0013 | Object Detection with HaarCascade | Detecting Faces with the HaarCascade Filter. |
| 0014 | Hand Tracking Volume Adjuster | Program that allows to adjust the speaker volume of your Laptop or Desktop using your hand movement. |
| 0015 | AI Attendence Marker using Face Recognition | Attendance marking system with a live feed. |
| 0016 | AI Curl Trainer | Curl Trainer built using a pose detector. |
| 0017 | Angle Finder | Angle Finder from a live feed. |
| 0018 | Annotating Images with Object Detection API | Annotating Images and Videos using multiple Object Detection models like SSDs, RCNNs, RFCNNs. |
| 0019 | Bar and QR Code Detection | Detecting Bar Codes and QR Codes from a live video feed and image. |
| 0020 | Document Scanner | Document Scanner from a live feed. |
| 0021 | Flickr8 Image Caption Generation | Image Captioner built using Flickr 8k Dataset. |
| 0022 | Space Radio Signals Classification | Space Radio Classification on SETI Data. |
| 0023 | Cell Instance Segmentation | Instance Segmentation on the Sartorius Cell Instance Segmentation Dataset. |
| 0024 | Google Street View House Number Recognition | House Number Recogition on the The Street View House Numbers (SVHN) Dataset. |
| 0025 | Object Dimension Measurer | Measure Object Dimensions from a live feed. |
| 0026 | Optical Mark Recognizer | Optical Mark Recognizer(OMR) from a live feed. |
| 0027 | Text Detector Tesseract | Text Detector built using Google's Tesseract Engine. |
| 0028 | Toxic Comments Classification | Toxic Comment Classification with 1D Convolutions. |
| 0029 | Virtual Keyboard | Virtual Keyboard that can be used with a live stream. |
| 0030 | Number Plate Detection with EfficientDet | Number Plate Recognition on the CCPD (Chinese City Parking Dataset, ECCV) using the EfficientDet Model. |
| 0031 | Wheat Detection with EfficientDet | Wheat Grain Detection using the EfficientDet Model. |
| 0032 | COTS Detection with EfficientDet | Crown-of-Thorns Starfish (COTS) Detection in the Great Barrier Reef using EfficientDet. |
| 0033 | SIIM-FISABIO-RSNA COVID-19 Detection With EfficientDet | Identifying and localizing COVID-19 abnormalities on chest radiographs using the EfficientDet Model. |
| 0034 | VinBigData Chest X-ray Abnormalities Detection With EfficientDet | Automatically localize and classify thoracic abnormalities from chest radiographs using the EfficientDet Model. |
| 0035 | NFL 1st and Future-Impact Detection With EfficientDet | Detect helmet impacts in videos of NFL plays using the EfficientDet Model. |
| 0036 | Oxford-IIIT Pet Dataset Classification With EfficientDet | Classification of pets into 37 categories using the EfficientDet Model. |
| 0037 | Aquarium Dataset Classification with EfficientDet | Classification of water species using the EfficientDet Model trained on the "Aquarium Dataset". |
| 0038 | Crowd Detection With EfficientDet | Crowd Counting/Detection using the EfficientDet Model. |
| 0039 | Herb Detection With EfficientDet | Herb Plants Detection using the EfficientDet Model. |
| 0040 | Swimming Pool Detection With EfficientDet | Swimming Pool Detection on Aerial Images using the EfficientDet Model. |
| 0041 | Face Mask Detection With EfficientDet | Face Mask localization and Detection using the EfficientDet Model. |
| 0042 | Tracheostomy Tubes Detection With EfficientDet | Tracheostomy Tubes localization and Detection on X-rays using the EfficientDet Model. |
| 0043 | Barcode Detection With EfficientDet | Barcode Detection using the EfficientDet Model. |
| 0044 | Gotukola Detection With EfficientDet | Detecting and Localizing Gotukola leaves using the EfficientDet Model. |
| 0045 | Helminths and Schistosoma Mansoni Detection With EfficientDet | Detecting and Localizing different species, such as roundworms, whipworms, and hookworms, which causes Soil-transmitted Helminths and Schistosoma Mansoni infections using the EfficientDet Model. |
| 0046 | Fire Detection With EfficientDet | Fire and Smoke Detection using the EfficientDet Model. |
| 0047 | Semantic Segmentation with AWS Sagemaker | Semantic Segmentation of Pet Data with AWS Sagemaker. |
| Paper ID | Title | Description |
|---|---|---|
| 0001 | EfficientDet :Scalable and Efficient Object Detection | EfficientDet is a type of object detection model, which utilizes several optimization and backbone tweaks, such as the use of a BiFPN, and a compound scaling method that uniformly scales the resolution,depth and width for all backbones, feature networks and box/class prediction networks at the same time. |
| 0002 | YOLACT: Real-time Instance Segmentation | YOLACT is a real-time instance segmentation system that is capable of detecting multiple objects and segmenting them in an image or a video stream. It achieves this by using a one-stage detector, which is a type of deep neural network that can detect objects in a single pass through the image, without requiring a separate region proposal step. The system also employs a fully convolutional mask head, which is another neural network that generates instance masks directly from the detector's output. By doing so, the system can perform instance segmentation in real-time while maintaining high accuracy. |
| 0003 | YOLACT++: Better Real-time Instance Segmentation | YOLACT++ is an improved version of the YOLACT (You Only Look At CoefficienTs) model, which is designed for real-time instance segmentation tasks. YOLACT++ uses a modified version of the ResNet-101 backbone architecture combined with a Feature Pyramid Network (FPN) and a Panoptic Feature Pyramid Network (PFPN) to extract features from different scales. It then uses a lightweight prototype projection head that generates mask coefficients for each instance, which are used to construct instance masks in real-time. |