Skip to content

Latest commit

 

History

History
83 lines (56 loc) · 3.48 KB

tasks.md

File metadata and controls

83 lines (56 loc) · 3.48 KB

100 Days of Computer Vision - Tasks

Days 1-6: Working with OpenCV in a Local Environment

Day 1: Camera Image Capture

  • Load an image from the camera and save it to the local directory.

Day 2: Image Loading and Resizing

  • Load an image from a file and resize it to a specific dimension.

Day 3: Loading and Grayscale Conversion

Day 4: Image Background Removal

  • Load the grayscale image from Day 3 and remove the image background.

Day 5: Drawing Circle on Live Camera Feed

  • Draw a circle on the live camera feed in real-time.

Day 6: Real-time Annotation on CV Camera Frame

  • Annotate your username and the date on the lower-left corner of the CV camera frame and display it in real-time.

Days 7-17: Advanced OpenCV Operations on Images

Day 7: Blob Analysis by Convexity

  • Perform Blob analysis using the convexity feature on an image from 'images/day6.jpg'.

Day 8: Canny Edge Detection on Live Video Feed

  • Apply Canny edge detection to the live video feed from the camera.

Day 9: OpenCV Averaging

  • Perform image smoothing using OpenCV averaging.

Day 10: OpenCV Median Blur

  • Implement the median blur operation using OpenCV.

Day 11: OpenCV Gaussian Blur

  • Apply Gaussian blur to an image using OpenCV.

Day 12: OpenCV Bilateral Filter

  • Implement a bilateral filter on an image using OpenCV.

Day 13: Adaptive Thresholding for Pencil Sketch

  • Utilize adaptive thresholding to create a pencil sketch effect on the live camera feed.

Day 14: Contours and Shape Detection

  • Implement contour detection and shape identification on any image. Refer to OpenCV documentation for implementation details.
  • Detect basic shapes like circles, rectangles, and triangles using contour properties and annotate them

Day 15: Mouse Event as a Paint Brush

  • Create a program that allows the mouse to be used as a paintbrush on a frame from the camera feed. Refer to OpenCV documentation for implementation details.

Day 16: Live Face Detection

Day 17: Facial Recognition Implementation

Days 18-20: Deep Learning Model Building and Hyperparameter Tuning

Day 18: Handwritten Digit Recognition with Keras

  • Load and display the first 20 digits from the Keras Handwriting MNIST dataset.
  • Include demo code to load the dataset.
from keras.datasets import mnist

# Load the MNIST dataset
(x_train, y_train), (x_test, y_test) = mnist.load_data()

# Display the first 20 images
# Add code snippet here

Day 19: Deep Learning Model Building on TensorFlow

  • Build a deep learning model with the architecture (32x64x128x32x10) using TensorFlow.
  • Train the model on the MNIST dataset from Day 18 and save the model.

Day 20: Model Hyperparameter Tuning for Improved Accuracy

  • Load the model built on Day 19 and perform hyperparameter tuning to improve the model's accuracy. Use grid search or random search techniques to find the best hyperparameters.