101 on improving OCR for Blurred Images | Blogs | nCircle Tech
Spread the word


OCR or optical character recognition has come a long way in the last decade. This technology provides a complete solution for form processing and document capturing. However, the process could harbour several distortions which results in poorly scanned photo/text-photo images and natural images, thus rendering the OCR unreliable. To combat this shortcoming, several methods supported by new age technologies have come into play over the last few years. We now have the ability to correct or remove the image distortions and improve the OCR accuracy to optimal levels, as per specific business needs.

While we have already seen the need for image processing, it would be delightful to know that there are several open source libraries available that will help you improve the optical character recognition accuracy. JAI media Apis, JMagick, ImageJ, AForge.Net, OpenCV, and ImageMagick are few of the renowned open source libraries that are capable of processing the images as per your needs.

Our Experience with ImageMagick:

We have tested the ImageMagick open source library while working on a 3D inspection based application and the results were phenomenal, especially when selecting and processing an object from huge inspection files.

ImageMagick allows users to create, edit, or convert images with support for over 200 types of files. Users can resize, flip, mirror, distort, rotate, and transform images along with a dozen other features. We used the program to improve the characters in the huge data files.

The crop function was highly useful in improving the accuracy of the text. Coupled with the sharpen function, this feature greatly improved the quality of the picture by sharpening the edges.

One unique feature of ImageMagick was the sampling tool. It allows us to take the samples from the image to adjust the variations due to noise in the picture. This ensures a high image quality with less noise, and much greater clarity for the viewer.

 Before and after image processing effects

Users also get additional features such as generalized pixel distortion, noise, and color reduction, transformation, and special effects. These features ensure that you get considerable improvement in the quality of the image. While these features can be utilized from the command-line, you can also use them from programs written in various languages such as C, C++, Pascal, Python, PHP, etc. The API and ABI also appear to be stable, mitigating any fears of a security breach. ImageMagick is capable of running on Linux, Windows, Mac Os X, iOS, Android OS, and others.

While not all agree on the benefits offered by image processing tools, practical use hints at tremendous advantages. One of the most common perks of the image processing tools is reducing the possibility of mistake or capturing wrong data. Users can also save the time of OCR and reduce efforts that otherwise would have to be invested in correcting the extracted data. Besides, processing the image before OCR ensures that words, text, tables, and data are identified according to the pre-set criteria of the software. It results in the categorization of data and graph clearly, thus enhancing the final output tremendously. 

Image courtesy: www,pexels.com

Authored by Sohel

Contact us

nCircle Tech (inCorporated in 2012) empowers passionate innovators to create impactful 3D visualization software for desktop, mobile and cloud. Our domain expertise in CAD-BIM customization driving automation with the ability to integrate advanced technologies like AI/ML and VR/AR; empowers our clients to reduce time to market and meet business goals. nCircle has a proven track record of technology consulting and advisory services for the AEC and Manufacturing industry across the globe. Our team of dedicated engineers, partner ecosystem and industry veterans are on a mission to redefine how you design and visualize.

Over the last 7+ years, the organisation has worked on more than 150 large and complex projects for 50+ customers across 15+ countries.

Spread the word