Fix detection prediction outputs

[sungchul2]

Summary

Ticket no. : 144396

Torch models have abnormal outputs, which seem unfiltered (especially ATSS and SSD).
It can yield wrong performance measurements by adding very many false positives.

box color : blue - GT | red - predictions

But, OV model have normal outputs because outputs are filtered before wrapped by DetectionResult

figure order : rtmdet_tiny OV Model | rtmdet_tiny Torch Model | atss_mobilenetv2 Torch Model

To fix this issue, this PR includes:

• Use confidence_threshold to filter outputs
  • model api uses it (default : 0.5) to filter outputs (link). To match the results, use confidence_threshold from checkpoint before wrapping outputs with DetBatchPredEntity.
    figure order : rtmdet_tiny OV Model | rtmdet_tiny Torch Model
    

How to test

Checklist

• I have added unit tests to cover my changes.​
• I have added integration tests to cover my changes.​
• I have ran e2e tests and there is no issues.
• I have added the description of my changes into CHANGELOG in my target branch (e.g., CHANGELOG in develop).​
• I have updated the documentation in my target branch accordingly (e.g., documentation in develop).
• I have linked related issues.

License

• I submit my code changes under the same Apache License that covers the project.
  Feel free to contact the maintainers if that's a concern.
• I have updated the license header for each file (see an example below).

# Copyright (C) 2024 Intel Corporation
# SPDX-License-Identifier: Apache-2.0