Error Level Analysis Tool
Detect image tampering with Error Level Analysis (ELA). Upload any JPEG or PNG image and instantly generate a heatmap revealing potential edited regions. Our tool re-saves the image at a known quality level and compares pixel differences - areas with inconsistent compression artifacts (shown in bright red) are likely to have been modified. Fully private, 100% browser-based, and free with no signup required.
Upload a JPEG or PNG image to detect edited regions using Error Level Analysis (ELA). The tool re-saves your image at a specific JPEG quality and highlights areas where compression artifacts differ - bright regions indicate possible tampering.
Drop an image here or click to browse
Supports JPEG and PNG images
How ELA works: The tool re-saves your image at a specific JPEG quality level, then compares pixel differences between the original and re-saved version. Unmodified areas compress uniformly (dark), while edited areas produce different compression artifacts (bright). All processing happens in your browser - no data is uploaded.
Why Use Our Error Level Analysis Tool?
Detect Image Tampering
Identify edited or modified regions in JPEG images by analyzing compression artifact inconsistencies. Bright areas on the heatmap indicate where the image has likely been altered.
Adjustable Sensitivity
Control JPEG re-save quality from 75% to 95% to balance detection sensitivity vs false positives. Lower quality amplifies differences for catching subtle edits.
Instant Browser-Based Analysis
All ELA processing runs locally in your browser using the Canvas API. Results appear instantly with no server uploads or queue waiting times.
100% Private & Secure
Your images never leave your device. Every pixel comparison, heatmap generation, and analysis happens locally - fully offline-capable and completely confidential.
When to Use Error Level Analysis
Photo Authentication
Verify whether a JPEG image has been edited or retouched by analyzing compression inconsistencies across the image surface.
Legal Evidence Verification
Check if submitted photographic evidence in legal cases has been tampered with. ELA provides a visual indicator of potential image manipulation.
Digital Forensics
Analyze images during forensic investigations to locate regions that may have been altered, composited, or airbrushed by perpetrators.
Journalism & Fact-Checking
Validate the authenticity of news images before publication. Detect potential manipulation in photos from unverified sources or anonymous submissions.
Insurance Fraud Detection
Analyze claim photos for signs of digital manipulation. Detect edited damage images, doctored receipts, or altered scene photographs.
Scientific Image Integrity
Verify that research images, microscopy photos, and experimental results have not been digitally altered or selectively edited.
Understanding Error Level Analysis (ELA)
What is Error Level Analysis?
Error Level Analysis (ELA) is a digital image forensics technique used to detect regions of an image that may have been digitally altered. It works by exploiting the way JPEG compression artifacts behave across an image. When a JPEG image is re-saved at a known quality level, unmodified regions compress predictably, while edited regions produce different compression patterns - revealing potential tampering.
How our ELA tool works
The tool loads your image onto an HTML Canvas, re-saves it as a JPEG at your chosen quality level, then compares every pixel between the original and the re-saved version pixel by pixel. Pixels with larger differences are displayed as brighter regions on the heatmap overlay. The tool reports the mean error, maximum error, and the percentage of significantly different pixels for quantitative analysis.
Interpreting ELA results
Understanding what the heatmap reveals requires some practice:
- Dark / uniform regions: Areas that are mostly black or very dark compress uniformly - these are likely unmodified parts of the original JPEG.
- Bright / red regions: Areas with bright red pixels indicate where the re-compressed JPEG differs significantly from the original - these are potential edit zones.
- Edge artifacts: Natural edges and sharp transitions (text, borders) may appear slightly brighter even without tampering, due to normal JPEG compression ringing.
- Mean error interpretation: A mean error below 3 typically indicates an unmodified image. Values above 10 suggest significant differences that warrant closer inspection.
Limitations & best practices
ELA is a powerful forensic indicator but not a definitive proof of tampering. For best results: use the original JPEG (not re-saved or transcoded images), test at multiple quality levels (75%, 85%, 95%) to confirm findings, combine ELA with other forensic techniques (metadata analysis, clone detection, file header analysis) for comprehensive verification, and be aware that heavy JPEG compression, noise reduction, or format conversion can produce false positive artifacts.
Related Tools
PNG to RAW
Convert PNG images to RAW format instantly - Free online PNG to RAW converter
PNG to DDS
Convert PNG to DDS texture format online free. Produces game-ready A8R8G8B8 DDS files for DirectX, Unity, and Unreal Engine - no upload, no signup.
PNG to TIFF
Convert PNG to TIFF online free. Lossless PNG to TIFF conversion in your browser - no upload, no signup required.
PNG to ICO
Convert PNG to ICO online free. Create multi-resolution favicons and Windows icons with frame sizes from 16×16 to 256×256 - no signup, no upload, browser-based.
Frequently Asked Questions
Error Level Analysis (ELA) is a digital forensics technique that identifies potential image tampering by analyzing JPEG compression artifacts. When a JPEG image is re-saved at a specific quality level, previously modified regions exhibit different compression characteristics compared to the original untouched areas, which are highlighted in the ELA heatmap output.
ELA is highly effective at detecting localized edits, especially copy-paste operations, airbrushing, and object removal in JPEG images. However, it is not 100% foolproof. Sophisticated editing in uncompressed formats (like PNG or TIFF), heavy noise reduction, or AI-generated content may produce less conclusive results. ELA should be used alongside other forensic methods for definitive conclusions.
Yes. The tool accepts both JPEG and PNG images. For PNG images, the tool first converts them to JPEG at the specified quality level before performing the ELA comparison. This means PNG images may show higher baseline error because of the initial conversion to JPEG compression.
The quality setting controls the JPEG compression level used when re-saving the image for comparison. Lower values (75-80%) create stronger compression artifacts, making differences more visible but potentially increasing false positives. Higher values (90-95%) are more conservative, showing only significant differences. The recommended default of 85% provides a good balance.
Natural image features like sharp edges, high-contrast text borders, and gradients can produce brighter ELA artifacts even in unedited images. JPEG compression inherently creates ringing artifacts around edges. Additionally, images that have been heavily compressed or resized may show uniform brightness patterns. Always compare multiple quality settings for reliable interpretation.
No. All processing happens entirely in your browser using the HTML Canvas API. Your image never leaves your device. The ELA heatmap is generated locally and you can choose to download it - nothing is stored, logged, or transmitted to any server.
ELA results can serve as a strong indicator of potential tampering, but they should not be used as the sole evidence in legal proceedings. For official use, always combine ELA with multiple forensic techniques and consult with a qualified digital forensics expert. The tool is designed for investigative guidance and educational purposes.
The tool supports JPEG (.jpg, .jpeg) and PNG (.png) image files. Large images are automatically resized to a maximum dimension of 2000px for analysis to maintain performance. There is no strict file size limit, but very large files may take longer to process in the browser.