Batch Image Skew and Four-Corner Correction Tool

Automatically detects the four corners of photographed documents, receipts, and whiteboards, batch corrects perspective and crops, and supports multiple document enhancements such as grayscale and binarization.

Bug report

Loading tool, please wait...

If the online tool fails to load or run, try the desktop tool.https://tools.yikeaigc.com/

Tool Usage

Return to old version

A desktop version is available with more powerful, stable features. Open desktop version

Instructions

Software Usage Instructions

Add images: Add images to be processed using the "Select File" or "Select Folder" buttons, or drag images or folders directly into the software window. Common formats such as JPG, PNG, BMP, TIFF, and WEBP are supported. Check "Traverse Subdirectories" to also process images in subfolders.
Select a scene preset: Choose a suitable preset based on the image type to quickly apply recommended parameters:
- General Documents: Suitable for most photographed documents and paper materials
- High-Definition Precision: Improves detection accuracy and enables enhancement, suitable for images with higher clarity requirements
- Receipts/Invoices: Optimized edge recognition for narrow documents such as receipts and invoices
- Whiteboard Enhancement: Suitable for photographed whiteboard and blackboard content, removing glare and shadows
- Black-and-White Scans: Outputs a black-and-white binarized effect, similar to scanner imaging
Set the output color mode: Choose original color, grayscale, or black-and-white document (binarization). In black-and-white mode, you can further select binarization methods such as adaptive, Otsu, or hybrid enhancement.
Configure enhancement options: Enable auto contrast, sharpening, denoising, shadow removal, white background stretching, and other processing as needed. You can also set parameters such as edge sensitivity, minimum document ratio, and processing size to adapt to different images.
Set output options: Choose the output format, JPEG quality, and blank margins, set the file name prefix/suffix, and choose how to handle cases where no document is detected (skip or save the original image), as well as the handling strategy for duplicate file names.
Specify the save directory and start: After selecting the output folder, click "Start Processing". The software will perform batch correction with multithreading. During processing, you can view the progress bar and real-time logs, and pause, resume, or stop the task.

FAQ

A: The tool performs edge detection and contour analysis on the image, then uses multiple criteria such as area ratio, aspect ratio, and centering to select the quadrilateral that most resembles a document as the four corners. It then applies perspective transformation to straighten it into a rectangular view. For images where the background and document have clear contrast, recognition is usually stable.

A: You can appropriately increase "Edge Sensitivity", adjust "Minimum Document Ratio" to match the document size in the frame, and increase "Processing Size" to retain more detail. When taking photos, keeping all four edges of the document visible and ensuring clear color contrast with the background will also help improve recognition accuracy. If it still cannot be recognized, you can select "Save Original Image (Do Not Crop)" in the settings to avoid missing any processing.

A: Original color preserves the image’s original colors and is suitable for documents with color content; grayscale converts it to black, white, and gray transitions, making the file smaller and easier to read; black-and-white document (binarization) processes the content into pure black and white, similar to a scanner effect, and is suitable for plain-text materials and archival storage.

A: No. All recognition, correction, and export operations are completed on your local device. Images will not be uploaded to any server, so you can safely process contracts, IDs, receipts, and other materials containing private information.

A: Yes. After selecting "Traverse subdirectories", the tool will automatically collect all images in the folder and its subfolders and process them in batches using multiple threads. After enabling "Keep original path structure", output files will be saved to the target folder according to the source directory hierarchy, making them easier to archive and find later.

A: You can choose a duplicate-name handling strategy in the settings. Auto rename appends a number to the file name to avoid conflicts; Skip existing keeps existing files and does not process them again; Overwrite existing replaces files with the same name with the new results. Auto rename is recommended by default to avoid accidental overwriting.

A: Recognition and correction use a certain amount of CPU and memory. Suggestions: 1) use the "Fast mode" preset and appropriately reduce the processing size; 2) reduce the number of threads running simultaneously; 3) process extremely large numbers of images in batches; 4) close other programs that use a lot of resources for a smoother processing experience.