fvilmos's repositories
tello_object_tracking
DJI Tello Drone tracks a moving object (a face or a person) on the horizontal/vertical plane keeps also a constant distance from the detection.
power_meter_mqtt
esp32 power meter for sensors with pulse output. Implements MQTT for status reporting. Accurate and easily can be integrated with Home Automation systems.
kinect_point_cloud
Draws real time 3D point cloud with kinect. The point cloud can be exported as .ply file, processed later with a viewer (included too).
object_classification_2d_to_3d
Images from an RGB-D camera are used to detect/classify objects in 2D, then detections are projected on the 3D point cloud.
cascade_tools
Automate the retraining of cascade classifiers in OpenCV
lane_detection
Lane detection with images captured from a front-facing camera in CARLA simulator.
shape_similarity_siamese_network
CNN based Siamese network for similarity detection. A toy database is used to demonstrate its functionality.
centroid_tracker
Simple centroid tracking algorithm using a cascade detector.
multi_task_learning_shape_color_position
Implementation of a slim multi-task learner DNN with 3 tasks (heads). Detects shape type, color, the position of the midpoint of the shape.
opencv_multi_graph
Simple but useful utility implementation for a multi-graph time series visualization in python with OpenCV.
road_segmentation
A small and fast CNN-based segmentation network is used to segment road / vehicles+pedestrians / other objects from the scene. The data collector uses CARLA simulator.
tello_serial
Translates Tello drone UDP commands to serial using an ESP-01.
crypto_coin_data_recorder
Application to record in an SQLite database the selected crypto coin values. The coin list is configurable, as same as the record frequency.
crypto_coin_telegram_bot
Track, inspect crypto coin-related information through a Telegram bot. Set alerts for selected crypto coins if a specific value (price or change in 24h) is reached.
perspective_transform_tool
Perspective transformation (birds eye) tool, for easy experimenting and finding the right transformation values.
shape_color_position_detection
A multi-headed DNN to detect shape type (circle, rectangle, triangle) shape color and position of the center point.
auto_pilot
Training, and testing a DNN model to drive a car in a virtual (urban) environment, using conditional imitation learning
AI_ReAct_agent
Implements a Reasoning and Acting AI agent with local LLM and tools usage in Python.
auto_annotator
Automatic data labeling tool, with flexible search patterns, using the power of GroundingDINO (detect anything with language).
cascade_nms
Non-Maximum-Suppression - with OpenCV cascade classifier. Uses Intersection Over Union to filter potential detections.
centernet_tiny
Resource optimal implementation of Centernet in Tensorflow-keras, for multi object detection.
color_filterer
Video stream color filterer (opencv)
frame_annotator
Frame (image) annotator tool, using mobile-SAM. Creates augmented data sets for Machine Learning model training.
gaussian_naive_bayes_classifier
Simple implementation of a Gaussian Naive Classifier in python
local_llm_utilities
use local LLMs to do Retrial-Augmented Generation with local Chroma, function calls and tool usage
machine_learning
Machine learning related work. Holds Kaggle competitions and other stuff
rapidly_exploring_random_tree
Simple implementation of a Rapidly-exploring Random Tree (RRT) algorithm for path planning.
ResNet_keras
ResNet implementation in Keras, alows a fexible way to generate the ResNet 18/34/50/101/152 or other model variants.
wifi_controlled_rc_car
control a cheap RC car over WiFi using HTTP GET commands, sent from a web browser or python file.