Abrar Istiak Akib's repositories
SORT-Multiple-Car-Tracking-using-YOLOv5
This is a Multi object tracker. Mainly it was build for tracking car movement in a junction.
TVL1_Optical_Flow_FrameGeneration
Generating optical flow frame by using TVL1 algorithm in organized way is not so difficult. But for a newbie it is toilsome to find this easier task. TVL1 is most popular way to generate optical flow frame which outperforms most other algorithm but lacks processing speed.
Synthetic-Tabular-Data-Generation-using-CTGAN-and-classify-with-XGboost
This is the repository to generate synthetic tabular data when the tabular data has imbalance in some feature.
ML-DL-others-Research-Paper-List
This is organized repo where i collect all my read or paper which i find interesting
Activity-Classification-from-First-Person-Office-Videos-with-Visual-Privacy-Protection
This is the official repository of the paper titled 'Activity Classification from First Person Office Videos with Visual Privacy Protection' which is accepted in IC4IR 2021.
Arithmetic-Logic-Unit-ALU-Design-and-Simulation-In-Verilog-and-Proteus
This is a repository for our EEE 304 course project. The name of the course is Digital Electronics Laboratory. We have explored the ALU section of modern CPU using the concept of elementary digital electronics. Here verilog HDL was coded using Quartus II 9.0 version software and 4 bit ALU hardware design was done using Proteus software.
Digital-Over-Voltage-Protection-System-for-Industrial-Loads-using-STM32-microcontroller
This is repository of our EEE 416 course project. In this repository, we organized our project files including all the necessary codes, simulation, PCB files.
Object-Based-Mask-Generation
Object based mask generator is used to generate mask from a video from dataset of Office activity recognition.Mainly the task of this is to differentiate the video from irrelevant background object so that network can concentrate into important objects to classify better.
Privacy_Protection_MASKRCNN_Resnet101
This is task 2 of VIP CUP 2019.Here we are providing security to sensitive objects while recognizing human activities captured by body cameras
Unsupervised_Anomaly_Detection_from-Multimodal_sensorial_info_of_Drone_by_Isolation-Forest
Unsupervised anomaly detection is now a days very vital thing in digital world.Anomaly always expected to happen rarely so unsupervised approach is necessary to deal with it.Isolation forest is one of the renowned method to detect anomaly unsupervised manner.Various use of isolation forest is showed in the repository.Anomaly detection from a heterogeneous autonomous system is mainly presented here.A multimodal sensorial information is used to detect anomaly and quantifying it in every timestamp.
Alzheimers-Stall-Catchers
Driven-Data Biomedical Research Competition
annotated_deep_learning_paper_implementations
🧑🏫 59 Implementations/tutorials of deep learning papers with side-by-side notes 📝; including transformers (original, xl, switch, feedback, vit, ...), optimizers (adam, adabelief, ...), gans(cyclegan, stylegan2, ...), 🎮 reinforcement learning (ppo, dqn), capsnet, distillation, ... 🧠
Gamma-Corrected-Video-Generation
When there is some challenge to overcome where it is the matter of brightness,luminosity then gamma correction works best among all pre-processing.To reproduce a good dataset over a less luminous dataset it may help.
Line-Follower
For the enthusiast who loves to build simple robot
MATLAB-Deep-Learning-Model-Hub
Discover pretrained models for deep learning in MATLAB
Microwave-Imaging
My research on Microwave Imaging.
Multiclass-Semantic-Segmentation-Annotation-tool-usage-codes
This is a repo to organize the tool use for multiclass semantic segmentation. Anyone who wishes to create a semantic segmentation for multiple classes can use this guideline. All the annotated image are saved in grayscale format.
Multilabel_Classification_of_Handwritten_BengaliGrapheme_Pytorch
It is training and testing repo for multiple output CNN.Here DenseNet of multiple variants have been implemented.It was created for the competition of 'Bengali.AI Handwritten Grapheme Classification' in Kaggle
Privacy_Protection_MaskRCNN_InceptionV2
This project was a part of our VIP CUP 2019 competition.It was task 2 of this year problem.Any privacy sensitive object should be preserved with privacy.
Robotics-Simulations-in-V-Rep-using-Matlab
Robotics simulation is one of the most important tasks in robotics. Pre-requisite of a robotics project is simulation. In the task of simulation, v-rep is a favorable choice as it has nice interactive environment. Moreover, v-rep is flexible with language. One can code in python, lua, or matlab etc. Here it is a repository where I keep all of my v-rep simulations in robotics.
Segment-Anything-Model-SAM-Point-prompt-Made-Easy-for-Segmentation-Binary-Mask-Evaluation
This is a made easy code for starters who want to use point prompt for recently released ground breaking foundational model Segment anything (SAM) by Meta AI. This code is to make easy to jump in a project where you can select point then apply to a directory and then evaluate on binary ground truth mask. You will also get binary mask from the code.
SPCUP2020-BUET-Synapticans
Proposed solution method for problem statement of IEEE Signal Processing Cup 2020 (Archived)
Symswarm-Symmetrical_Pattern_Formation_in_a_Swarm_of_Mobile_Robots
Symswarm is a symmetrical modular pattern formation algorithm which is for used to construct a symmetrical pattern among multiple mobile robots. The modular algorithm is divided into two phases where first phase is self guided independent phase and second phase is the swarm coordination phase between robots.
two-stream-action-recognition
Using two stream architecture to implement a classic action recognition method on UCF101 dataset