It is difficult to accurately size the defects that are oriented at an angle (that is not normal to the wave) using conventional amplitude based ultrasonic techniques. Since Time of Flight Diffraction (TOFD) is based on the diffraction of ultrasound at defect edges, defect sizing using this technique is amplitude independent. However, most of the TOFD based assessment relies on manual sizing, whose accuracy depends on quality of image and the operator's experience. Also, the utilization of TOFD for sections less than 15 mm has reportedly several difficulties. In this paper, we report our attempts to size the vertical and inclined defects using an in-house TOFD system built to inspect thin sections (6-10 mm). To improve sizing, automated defect sizing techniques termed Embedded Signal Identification Technique (ESIT) and Point Source Correlation Technique (PSCT) were developed. A ray tracing based model was also developed for a) optimizing the experimental parameters for thin sections, b) interpreting the received signals. Experiments were conducted on 10 mm thick samples with EDM defects and 6-7 mm welded maraging steel samples. The results obtained using manual and automated techniques were compared. Our comparisons lead us to believe that the automated defect sizing techniques can provide accurate and reliable results for thin sections. © 2005 American Institute of Physics.