Get more out of your images
Discover the features available in Dynamsoft Label Recognition SDK to take control of your data.
Extract and Recognize Characters from Various Image Formats
- Dynamsoft Label Recognition is capable of reading of alphanumeric characters and standard symbols, including period, commas, and dashes, of varying colours, font sizes and styles.
- Supported image formats include BMP, TIFF, JPEG, PNG and PDF.
Specify an Area to OCR Text Using a Reference Region
- Dynamsoft Label Recognition does not have to recognize texts from a full-sized image. Users have the option to run a zonal OCR feature for a text area. The OCR SDK allows you to specify the relative position of texts to a reference region.
- For example, a tag in a warehouse, where the texts of interest are always on the left and right of a barcode. The coordinates of the barcode can be used to specify the relative position of the text areas. Being able to specify this is useful because it further speeds scan recognition and improves recognition accuracy.
- Another example would be a price tag, where the text of interest is always on a yellow square background. In this case, the yellow square can serve as the reference region.
Use a Regular Expression to Improve Accuracy and Robustness
- Developers can also specify a Regular Expression to indicate to Dynamsoft Label Recognition the assumption about the texts. For example, an automotive VIN has a fixed number of digits and each digit has certain values to choose from. While Dynamsoft Label Recognition can usually scan all digits correctly as is, some challenging environments might pose obstacles. So, the semantics of a Regular Expression, to specify alphanumeric values that are expected to be scanned, can give Dynamsoft Label Recognition more context for better OCR results.
- After OCR finishes, a Regular Expression can also be used to verify and filter out texts that are not of interest. Developers can also use a character model to restrict alphanumeric or symbol possibilities. For example, you might specify numbers only, a mix of numbers and letters, specific fonts, special characters, and so on.
Sophisticated Image Pre-Processing Algorithms
- For images of poor quality, sophisticated algorithms built into Dynamsoft Label Recognition SDK are applied to improve contrast, remove noise, and so on. The characteristics of texts can also be used to improve the image quality. For example, with the algorithms the curve of the text can be used to correct distorted images.
Stitch Content Results from Neighboring Video Frames
- Dynamsoft Label Recognition SDK supports reading content from video streams. For a specified frame, the SDK extracts texts using OCR and cross references the result with neighbor frames for verifiable accuracy. The SDK can also intelligently decide if the text results are a partial or full match and overlap the partial results from sequential frames together.