Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,39 @@
|
|
| 1 |
---
|
| 2 |
license: cc0-1.0
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: cc0-1.0
|
| 3 |
---
|
| 4 |
+
# Material Recognition and Retrieval from images dataset
|
| 5 |
+
|
| 6 |
+
The goal of the dataset is zero-shot class agnostic material recognition and retrieval from images, for any material under any conditions.
|
| 7 |
+
Meaning the ability to recognize materials and their state from a single example in any domain.
|
| 8 |
+
|
| 9 |
+
This is fine-grained recognition of the same material and NOT classification of the material to specific set of classes.
|
| 10 |
+
It include recognizing same material, or same states of a material, base on similarity (aka which material or material states are more similar for example in case where material gradually transform like rotting apple).
|
| 11 |
+
|
| 12 |
+
This dataset is designed to address the general issue of one-shot material retrieval without
|
| 13 |
+
restrictions on material types, settings, and environments.
|
| 14 |
+
The main focus is on distinguishing between states of materials and identifying fine-grained categories such as rotten vs. ripe or coffee vs. cocoa.
|
| 15 |
+
|
| 16 |
+
The dataset is composed of two parts:
|
| 17 |
+
1) large scale synthetic train set focus on materials in general setting as well inside transparent container and materials with partial similarity and gradual transitions.
|
| 18 |
+
2) Real world natural images benchmarks for one-shot material recogntion focus on materials of of any type in any setting and on materials with partial similarity and gradual transitions.
|
| 19 |
+
3) atural image benchmark to test the ability to identify material states and subclasses using one or
|
| 20 |
+
a few examples (Figure 1).
|
| 21 |
+
|
| 22 |
+
# Paper:
|
| 23 |
+
|
| 24 |
+
[One-shot recognition of any material anywhere using contrastive learning with physics-based rendering
|
| 25 |
+
File Structure](https://openaccess.thecvf.com/content/ICCV2023/papers/Drehwald_One-Shot_Recognition_of_Any_Material_Anywhere_Using_Contrastive_Learning_with_ICCV_2023_paper.pdf)
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
# File Structure
|
| 29 |
+
|
| 30 |
+
Benchmark_MATSIM.zip: contain the benchmark made of real-world images as described in the paper
|
| 31 |
+
|
| 32 |
+
Dataset Generation Scripts.zip: Contain the Blender 4.1 Python scripts used for generating the dataset
|
| 33 |
+
|
| 34 |
+
MatSim_object_train_split_**.zip: Contain a subset of the synthetics dataset for images of CGI images materials on random objects as described in the paper.
|
| 35 |
+
|
| 36 |
+
MatSim_Vessels_Train_**.zip : Contain a subset of the synthetics dataset for images of CGI images materials inside transparent containers as described in the paper.
|
| 37 |
+
|
| 38 |
+
See .pdf files in for details on the dataset structure.
|
| 39 |
+
And .py files for readers.
|