It’s not for the preprocessed data, it’s for the postprocessed data. The inference from some models may output values that are out of image.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| How to get detections outside of the ROI with nvinfer | 10 | 177 | September 21, 2024 | |
| 0:00:04.100234080 2190201 0xfffefc053300 WARN | 12 | 637 | October 24, 2022 | |
| The inference works outside of the ROI (set by nvdspreprocess) | 3 | 493 | November 3, 2023 | |
| Infer the specified area data | 3 | 467 | October 12, 2021 | |
| Missing bounding boxes when using 'maintain-aspect-ratio=1' in nvinfer | 8 | 565 | September 23, 2022 | |
| Gst-nvdspreprocess set roi as input for pgie | 4 | 574 | June 27, 2023 | |
| Filtering based on Region of Interest (ROI) | 2 | 885 | October 12, 2021 | |
| Adding margin to ROI when inferencing on secondary GIE in Deepstream 5 | 4 | 565 | October 12, 2021 | |
| Deepstream preprocess roi | 14 | 283 | June 17, 2024 | |
| ROI for source video in deepstream-app | 14 | 4785 | October 12, 2021 |