https://github.com/skwzrd/imagesearch
OCR, EXIF, CLIP, ImageHash processing and search.
https://github.com/skwzrd/imagesearch
clip embedding exif exif-data-extraction flask image-processing image-recognition ocr ocr-recognition python search
Last synced: 5 months ago
JSON representation
OCR, EXIF, CLIP, ImageHash processing and search.
- Host: GitHub
- URL: https://github.com/skwzrd/imagesearch
- Owner: skwzrd
- Created: 2024-09-13T06:33:10.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2024-10-13T00:56:49.000Z (over 1 year ago)
- Last Synced: 2025-04-11T15:10:26.750Z (about 1 year ago)
- Topics: clip, embedding, exif, exif-data-extraction, flask, image-processing, image-recognition, ocr, ocr-recognition, python, search
- Language: Python
- Homepage:
- Size: 1.15 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# ImageSearch
A program that stiches together many libraries in order to provide a new image search experience.
Get your venv up with and install the `requirements.txt`.
- Run `python3.12 -m pip install git+https://github.com/openai/CLIP.git` to install CLIP.
- Run `python3.12 -m pip install -r requirements.txt`
- If you don't have nvidia drivers installed, you can use `sudo ubuntu-drivers autoinstall` and reboot.
Set your configs in `configs.py`.
Run `python3.12 process.py` to crawl images and populate a sqlite database with CLIP embeddings, OCR text, EXIF data, Hashes, Noise, and/or Faces.
Run `python3.12 web.py` to launch the web UI for searching the sqlite database.
## Dependencies
- For searching embeddings, we use [https://github.com/openai/CLIP](https://github.com/openai/CLIP) (see `requirements.txt` for install).
If you want OCR, pick one OCR program.
- ocrs (recommended) (download and compile [https://github.com/robertknight/ocrs](https://github.com/robertknight/ocrs))
- doctr (see `requirements.txt`)
- pytesseract (see `requirements.txt`)
If you want facial detection to search by face counts, install `face_recognition` (see `requirements.txt` ).
Need help choosing? See [https://neetventures.com/post/50](https://neetventures.com/post/50).
## Performance
My server (see below) does 160 images/second with EXIF, CLIP, and OCR ([https://github.com/robertknight/ocrs](https://github.com/robertknight/ocrs)) engaged.
## Hardware Requirements
These metrics were taken while processing images. I'll let you interpret them.
`sudo lshw -class CPU | grep -i product`
```
product: Intel(R) Core(TM) i5-9400F CPU @ 2.90GHz
```
`free -mh`
```
total used free shared buff/cache available
Mem: 31Gi 2.5Gi 20Gi 17Mi 8.0Gi 28Gi
```
`nvidia-smi`
```
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 550.107.02 Driver Version: 550.107.02 CUDA Version: 12.4 |
|-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce GTX 1660 Ti Off | 00000000:01:00.0 Off | N/A |
| 23% 41C P2 28W / 120W | 466MiB / 6144MiB | 5% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| 0 N/A N/A 25919 C python3.12 462MiB |
+-----------------------------------------------------------------------------------------+
```