Apple Detection from Fuji Dataset using Faster R-CNN -- 3
Budget: ₹1,500 – ₹12,500 INR
Multi model deep learning for Fuji Apple Detection
- use the KFuji RGB-DS dataset
- preprocess data including range correction of backscattered signal, 2D projection of 3D point cloud, and image registration between range-corrected intensity and raw RGB images
- use Faster R-CNN model for object detection to properly obtain ROI(region of interest) and the classification network. Input layer of original Faster R-CNN model is modified to accept 5-channel images. For this, filter from first layer increase from 3 to 5 in depth.
- Use anchor ratios (8,16,32) and
PS : project must be completed within 5 days. Message me for further details.
- use the KFuji RGB-DS dataset
- preprocess data including range correction of backscattered signal, 2D projection of 3D point cloud, and image registration between range-corrected intensity and raw RGB images
- use Faster R-CNN model for object detection to properly obtain ROI(region of interest) and the classification network. Input layer of original Faster R-CNN model is modified to accept 5-channel images. For this, filter from first layer increase from 3 to 5 in depth.
- Use anchor ratios (8,16,32) and
PS : project must be completed within 5 days. Message me for further details.