This study hypothesized that the absence of blurred images in the training sets of neural networks might lead them to overly depend on high spatial frequency information for object recognition, resulting in divergences from human visual processing. This hypothesis was systematically assessed by comparing different training regimes involving both clear and blurred images (i.e., standard training, weak blur training, and strong blur training). The results demonstrated that networks trained with blurred images surpassed standard networks in predicting neural responses to objects under diverse viewing conditions. Additionally, these blur-trained networks developed an increased sensitivity to object shapes and increased robustness to various types of visual disturbances, aligning more closely with human perceptual processes. Our research underscores the importance of incorporating blur as a vital component in the training process for neural networks to develop representations of the visual world that are more congruent with human perception. Based on our results, we recommend the integration of blur as a standard image augmentation technique in the majority of computer vision tasks. For those who are interested in the replication of the study or further exploration, the pretrained weights for 8 different networks are provided under the folder of Models. Additionally, the matlab code to generate the main figures in the manuscript are also provided under the folder of Figures. The training codes for our networks are available on our GitHub page at https://github.com/hojin89/BlurTraining. For any inquiries, please reach out to jangh@mit.edu.