Correlation Between Quality Evaluation Metrics and Teeth Detection Results in Panoramic X-Rays Using Deep Learning

Panoramic images are one of the most requested exams by dentists for allowing the visualization of the entire mouth. Interpreting X-ray images is a time-consuming task in which misdiagnoses can occur due to the inexperience or fatigue of professionals. In this work, we applied different image enhancement techniques as a pre-processing step to determine which image features correlate with improvements in teeth detection in panoramic images using deep learning architectures. We contrasted the performance of five object-detection architectures using 300 panoramic images of a public dataset. We evaluated the enhancement in the pre-processing step and the detection performance. Quality and detection metrics were considered, and the cross-correlation between them was computed for every object-detection method contemplated. We observe the dependence of the detection performance with some image enhancement techniques, especially those that introduce less noise and preserve the global contrast of the image.

Panoramic Dental Radiography Image Enhancement Using Multiscale Mathematical Morphology

Panoramic dental radiography is one of the most used images of the different dental specialties. This radiography provides information about the anatomical structures of the teeth. The correct evaluation of these radiographs is associated with a good quality of the image obtained. In this study, 598 patients were consecutively selected to undergo dental panoramic radiography at the Department of Radiology of the Faculty of Dentistry, Universidad Nacional de Asunción. Contrast enhancement techniques are used to enhance the visual quality of panoramic dental radiographs. Specifically, this article presents a new algorithm for contrast, detail and edge enhancement of panoramic dental radiographs. The proposed algorithm is called Multi-Scale Top-Hat transform powered by Geodesic Reconstruction for panoramic dental radiography enhancement (MSTHGR). This algorithm is based on multi-scale mathematical morphology techniques. The proposal extracts multiple features of brightness and darkness, through the reconstruction of the marker (obtained by the Top-Hat transformation by reconstruction) starting from the mask (obtained by the classic Top-Hat transformation). The maximum characteristics of brightness and darkness are added to the dental panoramic radiography. In this way, the contrast, details and edges of the panoramic radiographs of teeth are improved. For the tests, MSTHGR was compared with the following algorithms: Geodesic Reconstruction Multiscale Morphology Contrast Enhancement (GRMMCE), Histogram Equalization (HE), Brightness Preserving Bi-Histogram Equalization (BBHE), Dual Sub-Image Histogram Equalization (DSIHE), Minimum Mean Brightness Error Bi-Histogram Equalization (MMBEBHE), Quadri-Histogram Equalization with Limited Contrast (QHELC), Contrast-Limited Adaptive Histogram Equalization (CLAHE) and Gamma Correction (GC). Experimentally, the numerical results show that the MSTHGR obtained the best results with respect to the Contrast Improvement Ratio (CIR), Entropy (E) and Spatial Frequency (SF) metrics. This indicates that the algorithm performs better local enhancements on panoramic radiographs, improving their details and edges.

The influence of periapical lesions on the repeatability of two electronic apex locators in vivo

OBJECTIVES

To determine the influence of periapical lesions on the repeatability of two electronic apex locators (EALs), Root ZX II, and RomiApex A-15, in maxillary anterior teeth, in vivo.

MATERIALS AND METHODS

After approval by the local ethics committee, 43 maxillary anterior teeth scheduled for root canal therapy were assigned to two groups. The periapical lesion (PAL) group included 21 teeth with necrotic pulps and radiographically visible periapical lesion, and the normal periapex (NPA) group consisted of 22 teeth with vital pulps and no periapical lesion. In each canal, Root ZX II and RomiApex A-15 were used to determine working length defined as the "zero" reading on the display. Two consecutive measurements were performed with each EAL on each tooth and a second operator measured the endodontic file with a digital caliper without knowledge of the EAL being used or the tooth diagnosis. The absolute value of the difference between the first and second measurement (measurement pairs) was calculated. Repeatability was evaluated by the Bland-Altman method, and an aligned ranks transformation ANOVA was conducted to compare the EALs.

RESULTS

The median absolute difference between measurement pairs was significantly lower (p < 0.001) and the limits of agreement were narrower in the NPA group. Among the EALs, median absolute differences were significantly smaller for Root ZX (p < 0.001).

CONCLUSIONS

Measurements were more repeatable when no periapical lesion was present. Root ZX showed better repeatability than RomiApex.

CLINICAL RELEVANCE

Periapical lesions seem to influence successive in vivo measurements with EALs.