Method

UDeerDiffInstanceSegmentation [UDeer_DIS]
TBD

Submitted on 9 Aug. 2023 04:27 by
dong zichao (cuhk)

Running time:0.1 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
We use SOTA diffusion model as feature exactor.
Mask proposal mechanism is similar to Mask2former.
By designing custom PEFT method, we can improve our
model by only 200 annotated images.More details
would be updated later.
Parameters:
lr = 1e-5
Latex Bibtex:
TBD

Detailed Results

Currently, no detailed information available.

Test Set Average

AP AP50%
14.11 24.80
This table as LaTeX

Test Image 0

Input Image

Prediction


Test Image 1

Input Image

Prediction


Test Image 2

Input Image

Prediction


Test Image 3

Input Image

Prediction


Test Image 4

Input Image

Prediction


Test Image 5

Input Image

Prediction


Test Image 6

Input Image

Prediction


Test Image 7

Input Image

Prediction


Test Image 8

Input Image

Prediction


Test Image 9

Input Image

Prediction




eXTReMe Tracker