Face sketch recognition is a crucial field with applications in identifying suspects and criminals based on verbal descriptions (face sketches) provided by eyewitnesses. Although deep convolutional neural networks (DCNNs) have significantly advanced face recognition from photos, recognizing faces from sketches remains challenging due to texture differences and limited training samples. To overcome these challenges, an innovative methodology that integrates the shearlet transform as a pre-processing layer within the DCNN was proposed. This combination aims to establish a robust learning foundation for identifying individuals from face photos using their corresponding face sketches. Experimental evaluations showcase the effectiveness of our approach, achieving a remarkably high recognition rate. The incorporation of the shearlet transform enhances the DCNN's capability to handle texture disparities between face photos and sketches, resulting in improved performance. Our research marks the first instance of combining DCNN with the shearlet transform for face sketch recognition. Our approach proves highly effective in addressing sketch recognition challenges, as evidenced by an impressively low error rate of only 0.7%. This leads to minimized false positives, a crucial factor in law enforcement applications. A flawless recall score and an F1-score of 100% demonstrate exceptional performance in correctly identifying matches. This advancement carries promising implications for sensitive applications, such as recognizing suspects and criminals based on eyewitness descriptions, ultimately enhancing overall security and law enforcement efforts.