dhruvmsheth

Jetpack-Joyride

Clone of the Jetpack Joyride game written in C for Caltech's Software Design class

argus

Yet another vision-based pose estimator for the c u b e

ble-mask-detection-optimised

CTS-visual-drone-swarm-hackathon

Code for the paper "Collaborative Target Search with a Visual Drone Swarm: Adaptive Curriculum Embedded Multi-stage Reinforcement Learning Approach".

edge-impulse-cli

Installation of the Edge-Impulse-Cli on Ubuntu/Raspbian/Linux without requirement of administrator permissions or running into npm errors

Gelare

EdgeImpulse and Depth sensing powered LEGO Minstorms robot.

gen2-depthai-experiments

Experiments running Gen2 API, DepthAI on OAK-D. Alteration of existing examples, running models in sync

HapticCV

SpatialAI based Computer Vision for revolutionizing Bicycle Commute with Haptic Stimulus

Plant-Growth-Estimation-EdgeImpulse

self-sanitisation-person-detetction

This project has been designed for Arduino touch free-challenge Functioning: Usually, in malls or supermarkets, many people touch certain objects in aisles and it is really hard to monitor each object and sanitise each aisle individually. Most of the times, manual sanitisation is completed after each day and individual sanitisation of each aisle is not possible. Sometimes it happens that the food-aisle is more contaminated and the clothes-aisle is less contaminated, hence extra care is required for monitoring and sanitisatio of food-aisle. But such a monitoring solution is not possible manually. Therefore, I decided to make an autonomous solution based on TinyML deployed on Arduino Nano 33 BLE Sense. The person-identification model is highly accurate and uses an arducam which can be used in 5mp or 2mp variants. In solutions that contain ultrasonic sensors, or Lidar sensors, the readings or measurements are not accurate and can send false readings. Hence, the person detction model uses person recogition technology based on TensorflowLite framework. When, 1-50 people are detcted, a green light is switched on which means that the area is yet secure. When, 51-75 people are detected, a yellow light is flashed which means that awareness is required. When, 76-100 people are detected, a red light is flashed which means that people need to carefully touch objects and only touch those objects which are required. When 101-110 people are detected, the UV sanitisation system is activated which sanitises the complete area and objects are secure again. The count is again reset to 0 and the cycle is looped. The people vs Time graph is potrayed on the ThingSpeak web app where mall owners can track the time when the crowd is more or less and accordingly maintain more awareness during peak timings.

dhruvmsheth/gen2-depthai-experiments

Experiments running Gen2 API, DepthAI on OAK-D. Alteration of existing examples, running models in sync

dhruvmsheth/Plant-Growth-Estimation-EdgeImpulse

dhruvmsheth/Gelare

EdgeImpulse and Depth sensing powered LEGO Minstorms robot.

dhruvmsheth/Satellite-Image-Classification

dhruvmsheth/OpenCV_Competition2021

dhruvmsheth/argus

Yet another vision-based pose estimator for the c u b e

dhruvmsheth/CTS-visual-drone-swarm-hackathon

Code for the paper "Collaborative Target Search with a Visual Drone Swarm: Adaptive Curriculum Embedded Multi-stage Reinforcement Learning Approach".

dhruvmsheth/argus_depth

dhruvmsheth/cep

CEP is a software platform designed for users that want to learn or rapidly prototype using standard A.I. components.

dhruvmsheth/class

dhruvmsheth/cliport

CLIPort: What and Where Pathways for Robotic Manipulation

dhruvmsheth/dhruvmsheth.github.io

My personal webpage

dhruvmsheth/dhruvsheth-ai

dhruvmsheth/FoundationPose

[CVPR 2024 Highlight] FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects

dhruvmsheth/FutureGAN

Official PyTorch Implementation of FutureGAN

dhruvmsheth/HashNeRF-pytorch

Pure PyTorch Implementation of NVIDIA paper on Instant Training of Neural Graphics primitives: https://nvlabs.github.io/instant-ngp/

dhruvmsheth/isaac_ros_apriltag

NVIDIA-accelerated Apriltag detection and pose estimation.

dhruvmsheth/liquid_time_constant_networks

Code Repository for Liquid Time-Constant Networks (LTCs)

dhruvmsheth/misc

dhruvmsheth/misc-open

dhruvmsheth/neosensory-sdk-for-python

A Python SDK for interacting with Neosensory products. Powered by the Python Bleak library

dhruvmsheth/Oculaspectum

Code and Data for the research paper titled, "Exploring gaze behavior through object detection for walking pedestrians with Homonymous Hemianopia"

dhruvmsheth/PlantGAN

dhruvmsheth/point-cloud-prediction

Self-supervised Point Cloud Prediction Using 3D Spatio-temporal Convolutional Networks

dhruvmsheth/progressive_growing_of_gans

Progressive Growing of GANs for Improved Quality, Stability, and Variation

dhruvmsheth/Rigid-dynamics-gelare

Repository for exploring simulations and rigid dynamical systems to model constraints in Gelare.

dhruvmsheth/Spatio-Temporal-GAN

dhruvmsheth/spike-prime

Experiments with the LEGO Mindstorms (51515) and SPIKE Prime (45678)

dhruvmsheth/urc-misc

dhruvmsheth/visual-attention-oculaspecum

Predicting Human Eye Fixations via an LSTM-based Saliency Attentive Model. IEEE Transactions on Image Processing (2018)