A facial emotion prediction method based on ILTP

Abstract

The invention discloses a human face emotion pre-judgment method based on ILTP, which comprises the following steps of: making a human face emotion sample image library; extracting texture features ofa human face image by using an improved LTP algorithm ILTP for training; pre-judging the human face emotion by using the extracted texture features of the human face image; extracting the extracted texture features of the human face image by using the improved LTP algorithm ILTP for training; The method comprises the following steps: S11, collecting samples: collecting and sorting facial expression pictures; S12, sample classification: classifying the collected facial expression pictures. The invention is reasonably designed, easy to use, the improved image texture feature ILTP is used to extract face feature, the accuracy of facial expression classification depends on the validity of extracted features, Therefore, the efficient extraction and application of expression features is a key step to improve the expression recognition rate, At the same time, it combines (img file = '72567DEST_PATH_IMAGE002. TIF' wi= '33' he= '18' / ) algorithm to train and recognize face emotion samples, Finally, an algorithm fused with (img file = '52024DEST_PATH_IMAGE004. TIF' wi= '108' he= '24' / ) residual is used to improve the recognition rate. The invention is simple, convenient and suitable for widespread popularization.

Description

technical field

[0001] The invention relates to the technical field of facial emotion recognition and image processing, in particular to an ILTP-based human facial emotion prediction method. Background technique

[0002] Facial expression recognition is an important part of the research on artificial emotion and artificial psychology, involving various disciplines such as physiology, psychology, image processing, pattern recognition, machine vision, computer graphics, artificial intelligence, and cognitive science. application prospects. For example, natural and harmonious human-computer interaction, safe driving, intelligent monitoring, identity verification, intelligent lie detection, medical monitoring, behavioral science, psychological research, psychoanalysis, etc. Facial expression recognition plays an important role in harmonious human-computer interaction. In-depth research on expression recognition can not only make computers better understand human emotions and ps...

Examples

Embodiment 1

[0123] A facial emotion prediction method based on ILTP, including making a facial emotion sample library, using the improved LTP algorithm ILTP to extract the texture features of the face image, and using the extracted texture features of the face image to predict the facial emotion;

[0124] Described making facial emotion sample library comprises the following steps:

[0125] S11. Sample collection: collecting and arranging facial expression pictures;

[0126] S12, sample classification: classify and process the collected facial expression pictures; divide facial emotions into three categories: neutral, positive and negative; neutral emotions are people's facial expressions under normal circumstances; positive emotions are people's facial expressions when they are happy Facial expressions; negative emotions include anger, anger, sadness, surprise, and negative emotions;

[0127] S13. Sample normalization: perform an image size conversion operation on the original sample im...

Embodiment 2

[0143] A facial emotion prediction method based on ILTP, including making a facial emotion sample library, using the improved LTP algorithm ILTP to extract the texture features of the face image, and using the extracted texture features of the face image to predict the facial emotion;

[0144] Described making facial emotion sample library comprises the following steps:

[0145] S11. Sample collection: collecting and arranging facial expression pictures;

[0146] S12, sample classification: classify and process the collected facial expression pictures; divide facial emotions into three categories: neutral, positive and negative; neutral emotions are people's facial expressions under normal circumstances; positive emotions are people's facial expressions when they are happy Facial expressions; negative emotions include anger, anger, sadness, surprise, and negative emotions;

[0147] S13. Sample normalization: perform an image size conversion operation on the original sample im...

technical field

[0001] The invention relates to the technical field of facial emotion recognition and image processing, in particular to an ILTP-based human facial emotion prediction method. Background technique

[0002] Facial expression recognition is an important part of the research on artificial emotion and artificial psychology, involving various disciplines such as physiology, psychology, image processing, pattern recognition, machine vision, computer graphics, artificial intelligence, and cognitive science. application prospects. For example, natural and harmonious human-computer interaction, safe driving, intelligent monitoring, identity verification, intelligent lie detection, medical monitoring, behavioral science, psychological research, psychoanalysis, etc. Facial expression recognition plays an important role in harmonious human-computer interaction. In-depth research on expression recognition can not only make computers better understand human emotions and ps...

Examples

Embodiment 1

[0123] A facial emotion prediction method based on ILTP, including making a facial emotion sample library, using the improved LTP algorithm ILTP to extract the texture features of the face image, and using the extracted texture features of the face image to predict the facial emotion;

[0124] Described making facial emotion sample library comprises the following steps:

[0125] S11. Sample collection: collecting and arranging facial expression pictures;

[0126] S12, sample classification: classify and process the collected facial expression pictures; divide facial emotions into three categories: neutral, positive and negative; neutral emotions are people's facial expressions under normal circumstances; positive emotions are people's facial expressions when they are happy Facial expressions; negative emotions include anger, anger, sadness, surprise, and negative emotions;

[0127] S13. Sample normalization: perform an image size conversion operation on the original sample im...

Embodiment 2

[0143] A facial emotion prediction method based on ILTP, including making a facial emotion sample library, using the improved LTP algorithm ILTP to extract the texture features of the face image, and using the extracted texture features of the face image to predict the facial emotion;

[0144] Described making facial emotion sample library comprises the following steps:

[0145] S11. Sample collection: collecting and arranging facial expression pictures;

[0146] S12, sample classification: classify and process the collected facial expression pictures; divide facial emotions into three categories: neutral, positive and negative; neutral emotions are people's facial expressions under normal circumstances; positive emotions are people's facial expressions when they are happy Facial expressions; negative emotions include anger, anger, sadness, surprise, and negative emotions;

[0147] S13. Sample normalization: perform an image size conversion operation on the original sample im...

Images