Method for extracting a carious lesion area

Inventors

Yan, Jiayong • Wang, Wei • Pan, Liangliang • Wong, Victor C.

Assignees

Carestream Dental LLC

Interested in licensing this patent?

MTEC can help explore whether this patent might be available for licensing for your application.

Publication Number

US-8908936-B2

Patent

Publication Date

2014-12-09

Expiration Date


Abstract

A method for extracting a carious lesion area from sound regions of a tooth. In one embodiment, the method includes generating a digital image of the tooth; identifying tooth regions; extracting a suspicious lesion area by using a marker-controlled watershed algorithm or by using a morphological bottom-hat based method along with a multi-resolution surface reconstruction method; and removing false positives. In another embodiment, the method includes generating a digital image of the tooth comprising obtaining a fluorescence image of the tooth, obtaining a reflectance image of the tooth, and combining image data for the fluorescence and reflectance images; identifying tooth regions; extracting suspicious lesion areas; and removing false positives.

Core Innovation

A method is provided for extracting one or more lesion areas from sound regions of a tooth from a digital image of the tooth. The method accesses the digital image of the tooth and identifies one or more tooth regions by processing the digital image of the tooth. The identified tooth regions are obtained by generating a first threshold image from a grayscale fluorescence image and a second threshold image from a grayscale reflectance image, and forming a preliminary tooth regions image from their intersection. A reference binary image is generated from the grayscale fluorescence image using a third threshold value that exceeds the first threshold value, and a refined tooth regions image is generated from regions in the preliminary tooth regions image that are connected to objects in the reference binary image.

Suspicious lesion areas are extracted by applying either a marker-controlled watershed algorithm or a morphological bottom-hat based method along with a multi-resolution surface reconstruction method. In the marker-controlled watershed approach, internal markers and external markers are used, and a gradient image is processed to support the marker-controlled watershed transformation. In the morphological bottom-hat approach, bottom-hat based processing is carried out together with multi-resolution surface reconstruction, and lesion-area candidates are identified from the resulting processing.

False positives are identified and removed from the extracted suspicious lesion areas. False-positive removal includes identifying and removing one or more interproximal false positives by locating one or more interproximal regions and removing the interproximal false positives. The interproximal regions can be identified using a distance transformation and internal and external markers for marker-controlled watershed processing, and by identifying at least one interproximal region as pixels having substantially the same distance to two groups of water basins.

Claims Coverage

Two independent claims are present. The claims cover fluorescence and reflectance processing for identifying tooth regions, extraction of suspicious lesion areas using marker-controlled watershed or morphological bottom-hat with multi-resolution surface reconstruction, and identification and removal of false positives.

Dual-modality fluorescence and reflectance threshold refinement for tooth regions

Identifying one or more tooth regions by generating a first threshold image from a grayscale fluorescence image, generating a second threshold image from a grayscale reflectance image, forming a preliminary tooth regions image from the intersection of the first and second threshold images, generating a reference binary image from the grayscale fluorescence image using a third threshold value exceeding the first threshold value, and generating a refined tooth regions image from regions in the preliminary tooth regions image that are connected to objects in the reference binary image.

Suspicious lesion extraction via marker-controlled watershed or morphological bottom-hat with multi-resolution surface reconstruction

Extracting one or more suspicious lesion areas by applying a marker-controlled watershed algorithm or by applying a morphological bottom-hat based method along with a multi-resolution surface reconstruction method.

Interproximal false-positive detection and removal

Identifying and removing one or more false positives from the extracted one or more suspicious lesion areas.

Lesion extraction using fluorescence/reflectance combination and morphological bottom-hat with multi-resolution surface reconstruction

Generating a digital image comprising obtaining a fluorescence image, obtaining a reflectance image, and combining image data for the fluorescence and reflectance images; identifying one or more tooth regions; extracting one or more suspicious lesion areas using a morphological bottom-hat based method along with a multi-resolution surface reconstruction method; and identifying and removing one or more false positives from the extracted suspicious lesion areas.

The claims combine fluorescence and reflectance image processing, suspicious lesion extraction, and false-positive removal, including interproximal false positives.

Stated Advantages

Documented Applications

No documented applications found

JOIN OUR MAILING LIST

Stay Connected with MTEC

Keep up with active and upcoming solicitations, MTEC news and other valuable information.