File size: 1,293 Bytes
13b4edf
408a1cf
13b4edf
 
 
 
 
 
 
 
16303c6
13b4edf
0add1e3
0b63707
16303c6
86dcfea
16303c6
0b63707
16303c6
3b1b473
b6f2026
 
4873618
eff1ba3
4873618
eff1ba3
4873618
 
b6f2026
eff1ba3
4873618
8ebf724
16303c6
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
---
title: Amazon Products Demo
emoji: 📈
colorFrom: indigo
colorTo: pink
sdk: gradio
sdk_version: 3.41.2
app_file: app.py
pinned: false
---
Description:

We present a demo for performing object segmentation using a model trained on Amazon's ARMBench dataset. The model was trained on over 37,000 training images and validated on 4,425 images.

Usage:

You can use our demo by uploading your product image, and it will provide you with a segmented image.

Dataset:

-The model was trained on the ARMBench segmentation dataset, which comprises more than 50,000 images. 

-License: Creative Commons

-Paper: ARMBench: An object-centric benchmark dataset for robotic manipulation

-Authors: Chaitanya Mitash, Fan Wang, Shiyang Lu, Vikedo Terhuja, Tyler Garaas, Felipe Polido, Manikantan Nambi


You can learn more about this dataset on https://www.amazon.science/blog/amazon-releases-largest-dataset-for-training-pick-and-place-robots.


Download Dataset:

To download the dataset we used, you can use the following command in colab :

!wget https://armbench-dataset.s3.amazonaws.com/segmentation/armbench-segmentation-0.1.tar.gz


Feel free to explore and use this repository for your object segmentation needs. If you have any questions or need assistance, please don't hesitate to reach out