📝 Pipeline

*** We have the following 7 steps in our pipeline:
$ prepare_images
$ compute_keypoints_descriptors
$ image_matching
$ data_feature_matching
$ compute_k_matrix
$ generate_point_cloud
$ 3D reconstruction

🏛️ Datasets

• snow-man.
• hammer.
• cottage

? Something simplier and on our own terms

❌ Problems

Annoying Magic numbers

• number of clusters while image matching.
• Ransic threshold.

Triangulation Step looks to UnFold Object Points

Masking Regular Objects for datasets with UnMasked Images

🧐 What we noticed that caused poor feature match:

• Blury images.
• Change in focus.
• Change in distance.

❓ Questions

Triangulation ?

📊 Figures and Plots

Snow-Man Points Cloud

Hammer Feature Matches

Good Feature Match	Bad Feature Match

Hammer Feature Match Animated Video

output_1_2.mp4

Triangulation With Hand-Written Labels (Yesterday)

Traingulation Without Hand-Written Labels (Yesterday)

License

This project is licensed under the terms of the GNU General Public License version 3.0 (GPLv3). See the LICENSE file for details.