> For the complete documentation index, see [llms.txt](https://jgoodman8.gitbook.io/iron-data-science-notebook/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://jgoodman8.gitbook.io/iron-data-science-notebook/ml-datascience/computer-vision/object-detection/two-stage-detectors/r-cnn.md).

# R-CNN

Compared to the generic traditional pipeline, the feature extraction is performed using a CNN. This way:

1. Selective Search produces 2,000 region proposals
2. The CNN model extracts a 4,096 vector for each region.
3. An SVM classifies the region into background or object classes.

<figure><img src="https://569842953-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-LmjpNbCRLUyGiAxD8kn%2Fuploads%2FDDTyAXi0LGpsw9mVDKED%2Fimage.png?alt=media&amp;token=4b0fbb79-4dd3-47d6-adc9-549d4cfb599e" alt=""><figcaption></figcaption></figure>

**Drawbacks:**

1. Multi-stage model ->  cannot be trained end-2-end.
2. Uses disk caching of the extracted features -> exponential growth of disk usage.
3. Selective Search is slow.
4. The CNN is run on each independent region -> very slow