Henri Maître is Emeritus Professor at Télécom-Paris in France and was director of research at Télécom-Paris and the LTCI laboratory. He specializes in image processing and pattern recognition.

• Introduction: Image and Gaze xiChapter 1 The Legacy of Philosophers 11.1. The objectivist approach 31.1.1. The source: ancient Greece 31.1.2. After Greece 51.1.3. Kant and modern aesthetics 71.1.4. Objectivism after Kant: from pseudo-subjectivism to aesthetic realism 91.2. The subjectivist approach 131.2.1. From classicism to romanticism 141.2.2. The moderns 151.2.3. The influence of neurobiology 181.3. Subjectivism and objectivism: an ongoing debate 19Chapter 2. Neurobiology or the Arbitrator of Consciousness 252.1. fMRI protocols and neuroaesthetics 272.2. The fMRI quest for “beauty processes” in the brain 282.2.1. The role of the prefrontal cortex 282.2.2. The role of the insular cortex 312.2.3. The role of the visual areas 332.2.4. The role of memory and cognition 352.2.5. The role of embodiment 352.3. Responses from functional electric encephalography 362.4. A global cognitive scheme for aesthetic judgment? 392.4.1. J. Petitot’s neurogeometric model 402.4.2. A. Chatterjee’s aesthetic emotion model 402.4.3. The model by Brown et al. 422.4.4. Model proposed by H. Leder 432.4.5. The model by C. Redies 452.4.6. The emotions model developed by S. Koelsch et al. 472.4.7. L.H. Hsu’s model of emotions based on A. Damásio 472.4.8. Other models 502.5. A critique of neuroaesthetic methods 512.5.1. Criticism of neuroaesthetic methods 512.5.2. Criticisms of the objectives of neuroaesthetics 52Chapter 3. What Are the Criteria For a Beautiful Photo? 553.1. Before we enter into the fray 563.1.1. What reference books do we have? 563.1.2. “Beauty of an image” or “quality of an image”? 573.1.3. A glossary of aesthetic appraisal 583.1.4. Measuring beauty 603.2. Composition 633.2.1. Complexity versus simplicity 633.2.2. Unity 643.2.3. A specific case in composition: landscapes 643.2.4. Using oculometry to analyze composition 673.2.5. Format or aspect ratio 683.2.6. The rule of thirds (RoT) 703.2.7. The center of the image 723.2.8. Other rules for composition 733.3. Histograms, spectral properties and textures 763.3.1. Histograms and gray levels 763.3.2. Focus, spectral density, fractals 783.3.3. Textures 803.4. Color 823.4.1. About the concept of color 823.4.2. Preferences related to isolated colors 843.4.3. Preferences related to color palettes 863.5. What behavioral psychosociology has to say 933.5.1. Images of nature 933.5.2. The aesthetics of faces 963.5.3. The role of the signature, title and context 993.5.4. Perception and memory: prototypicality 101Chapter 4. Algorithmic Approaches to “Calculate” Beauty 1034.1. First steps: C. Henry 1034.2. G.D. Birkhoff’s mathematical approach 1044.3. Those who followed G.D. Birkhoff 1064.3.1. Beauty according to H.J. Eysenck 1064.3.2. The Post-War years: the designers, A. Moles and M. Bense 1064.3.3. A dynamic approach: P. Machado and A. Cardoso 1074.3.4. Work carried out by J. Rigau, M. Feixas and M. Bert 1084.4. Algorithmic approach with AI: J. Schmidhuber 110Chapter 5. The Holy Grail of the Digital World: Artificial Intelligence 1135.1. Which artificial intelligence? 1145.1.1. The principles 1145.1.2. Learning algorithms 1155.2. Why artificial intelligence in aesthetics? 1165.3. Expert opinions 1185.4. The database 1205.4.1. Generalist databases, used for aesthetic judgments 1225.4.2. Databases that are specialized for aesthetic photography 1265.4.3. Databases dedicated to artistic judgment 1295.4.4. Other image databases that are sometimes used 1305.4.5. Increasing databases 131Chapter 6. Primitive-based Classification Methods 1336.1. Judging aesthetics 1366.1.1. Multimedia primitives: the ACQUINE system (Datta et al.) 1366.1.2. Edges and chromatic distance: Ke et al. 1376.1.3. Photography rules: Luo and Tang and Mavridaki and Mezaris 1406.1.4. High-level primitives: Dhar et al. 1436.1.5. Generic descriptors of vision: Marchesotti et al. 1446.2. Help in composing beautiful photos 1486.2.1. The library of aesthetic primitives developed by Su et al. 1486.2.2. The OSCAR system by Yao et al. 1486.2.3. Embedded systems: Lo et al. and Wang et al. 1506.3. Some specific research related to the evaluation of aesthetics using primitives 1516.3.1. Color harmony: Lu et al. 1516.3.2. Group photography: Wang et al. 1536.3.3. Social networks and crowdsourcing: Schifanella et al. 1536.3.4. Looking at comments: San Pedro et al. 154Chapter 7. Deep Neural Network Systems 1557.1. DNNs dedicated to aesthetic evaluation 1577.1.1. High and low resolutions: the RAPID system, Lu et al. 1577.1.2. The multi-path DMA-Net architecture: Lu et al. 1607.1.3. Adapting to the size of the image: Mai et al. 1607.1.4. Finding beauty on the Web: Redi et al. 1637.1.5. Siamese and GAN networks: Kong et al. and Deng et al. 1657.1.6. Paying attention to the image construction: A-Lamp 1677.2. Variants around the basic DNN architecture 1707.2.1. Comparing photos between themselves: Schwarz et al. 1707.2.2. Making use of knowledge of the subject: Kao et al. 1727.2.3. BDN: halfway between classification and DNN 1747.2.4. Using the distribution of the evaluations 1757.2.5. Extracting a “dramatic” image from a panorama: the Creatism system 1787.3. Written appraisals: analyzing them and formulating new ones 1797.3.1. Photo critique captioning dataset (PCCD) 1817.3.2. Neural aesthetic image retriever (NAIR) 1827.3.3. Semantic processing by Ghosal et al. 1827.3.4. Aesthetic multi attribute network (AMAN) 1837.4. Measuring subjective beauty 1857.4.1. Recommendation systems 1867.4.2. Defining the user’s psychological profile 1887.4.3. Learning the user’s tastes through tests 1917.4.4. Multiplying concurrent expertise 194Chapter 8. A Critical Analysis of Machine Learning Techniques 1978.1. The popularity of studies on aesthetics 1978.2. A summary of learning methods 1998.2.1. Which architecture? Which software? 1998.2.2. What performances? 2008.3. Questioning the hypotheses 2038.4. Specific features of beautiful images detected by a computer 2048.4.1. Some observations on the photos in the AVA database 2058.4.2. The scores in the AVA database 207Conclusion 213Appendix 1. A Brief Review of Aesthetics 219Appendix 2. Aesthetics in China 237Appendix 3. The Aesthetic of Persian Miniatures 251Appendix 4. Aesthetics in Japan 263References 271Index 295

Henri Maître is Emeritus Professor at Télécom-Paris in France and was director of research at Télécom-Paris and the LTCI laboratory. He specializes in image processing and pattern recognition.

• Introduction: Image and Gaze xiChapter 1 The Legacy of Philosophers 11.1. The objectivist approach 31.1.1. The source: ancient Greece 31.1.2. After Greece 51.1.3. Kant and modern aesthetics 71.1.4. Objectivism after Kant: from pseudo-subjectivism to aesthetic realism 91.2. The subjectivist approach 131.2.1. From classicism to romanticism 141.2.2. The moderns 151.2.3. The influence of neurobiology 181.3. Subjectivism and objectivism: an ongoing debate 19Chapter 2. Neurobiology or the Arbitrator of Consciousness 252.1. fMRI protocols and neuroaesthetics 272.2. The fMRI quest for “beauty processes” in the brain 282.2.1. The role of the prefrontal cortex 282.2.2. The role of the insular cortex 312.2.3. The role of the visual areas 332.2.4. The role of memory and cognition 352.2.5. The role of embodiment 352.3. Responses from functional electric encephalography 362.4. A global cognitive scheme for aesthetic judgment? 392.4.1. J. Petitot’s neurogeometric model 402.4.2. A. Chatterjee’s aesthetic emotion model 402.4.3. The model by Brown et al. 422.4.4. Model proposed by H. Leder 432.4.5. The model by C. Redies 452.4.6. The emotions model developed by S. Koelsch et al. 472.4.7. L.H. Hsu’s model of emotions based on A. Damásio 472.4.8. Other models 502.5. A critique of neuroaesthetic methods 512.5.1. Criticism of neuroaesthetic methods 512.5.2. Criticisms of the objectives of neuroaesthetics 52Chapter 3. What Are the Criteria For a Beautiful Photo? 553.1. Before we enter into the fray 563.1.1. What reference books do we have? 563.1.2. “Beauty of an image” or “quality of an image”? 573.1.3. A glossary of aesthetic appraisal 583.1.4. Measuring beauty 603.2. Composition 633.2.1. Complexity versus simplicity 633.2.2. Unity 643.2.3. A specific case in composition: landscapes 643.2.4. Using oculometry to analyze composition 673.2.5. Format or aspect ratio 683.2.6. The rule of thirds (RoT) 703.2.7. The center of the image 723.2.8. Other rules for composition 733.3. Histograms, spectral properties and textures 763.3.1. Histograms and gray levels 763.3.2. Focus, spectral density, fractals 783.3.3. Textures 803.4. Color 823.4.1. About the concept of color 823.4.2. Preferences related to isolated colors 843.4.3. Preferences related to color palettes 863.5. What behavioral psychosociology has to say 933.5.1. Images of nature 933.5.2. The aesthetics of faces 963.5.3. The role of the signature, title and context 993.5.4. Perception and memory: prototypicality 101Chapter 4. Algorithmic Approaches to “Calculate” Beauty 1034.1. First steps: C. Henry 1034.2. G.D. Birkhoff’s mathematical approach 1044.3. Those who followed G.D. Birkhoff 1064.3.1. Beauty according to H.J. Eysenck 1064.3.2. The Post-War years: the designers, A. Moles and M. Bense 1064.3.3. A dynamic approach: P. Machado and A. Cardoso 1074.3.4. Work carried out by J. Rigau, M. Feixas and M. Bert 1084.4. Algorithmic approach with AI: J. Schmidhuber 110Chapter 5. The Holy Grail of the Digital World: Artificial Intelligence 1135.1. Which artificial intelligence? 1145.1.1. The principles 1145.1.2. Learning algorithms 1155.2. Why artificial intelligence in aesthetics? 1165.3. Expert opinions 1185.4. The database 1205.4.1. Generalist databases, used for aesthetic judgments 1225.4.2. Databases that are specialized for aesthetic photography 1265.4.3. Databases dedicated to artistic judgment 1295.4.4. Other image databases that are sometimes used 1305.4.5. Increasing databases 131Chapter 6. Primitive-based Classification Methods 1336.1. Judging aesthetics 1366.1.1. Multimedia primitives: the ACQUINE system (Datta et al.) 1366.1.2. Edges and chromatic distance: Ke et al. 1376.1.3. Photography rules: Luo and Tang and Mavridaki and Mezaris 1406.1.4. High-level primitives: Dhar et al. 1436.1.5. Generic descriptors of vision: Marchesotti et al. 1446.2. Help in composing beautiful photos 1486.2.1. The library of aesthetic primitives developed by Su et al. 1486.2.2. The OSCAR system by Yao et al. 1486.2.3. Embedded systems: Lo et al. and Wang et al. 1506.3. Some specific research related to the evaluation of aesthetics using primitives 1516.3.1. Color harmony: Lu et al. 1516.3.2. Group photography: Wang et al. 1536.3.3. Social networks and crowdsourcing: Schifanella et al. 1536.3.4. Looking at comments: San Pedro et al. 154Chapter 7. Deep Neural Network Systems 1557.1. DNNs dedicated to aesthetic evaluation 1577.1.1. High and low resolutions: the RAPID system, Lu et al. 1577.1.2. The multi-path DMA-Net architecture: Lu et al. 1607.1.3. Adapting to the size of the image: Mai et al. 1607.1.4. Finding beauty on the Web: Redi et al. 1637.1.5. Siamese and GAN networks: Kong et al. and Deng et al. 1657.1.6. Paying attention to the image construction: A-Lamp 1677.2. Variants around the basic DNN architecture 1707.2.1. Comparing photos between themselves: Schwarz et al. 1707.2.2. Making use of knowledge of the subject: Kao et al. 1727.2.3. BDN: halfway between classification and DNN 1747.2.4. Using the distribution of the evaluations 1757.2.5. Extracting a “dramatic” image from a panorama: the Creatism system 1787.3. Written appraisals: analyzing them and formulating new ones 1797.3.1. Photo critique captioning dataset (PCCD) 1817.3.2. Neural aesthetic image retriever (NAIR) 1827.3.3. Semantic processing by Ghosal et al. 1827.3.4. Aesthetic multi attribute network (AMAN) 1837.4. Measuring subjective beauty 1857.4.1. Recommendation systems 1867.4.2. Defining the user’s psychological profile 1887.4.3. Learning the user’s tastes through tests 1917.4.4. Multiplying concurrent expertise 194Chapter 8. A Critical Analysis of Machine Learning Techniques 1978.1. The popularity of studies on aesthetics 1978.2. A summary of learning methods 1998.2.1. Which architecture? Which software? 1998.2.2. What performances? 2008.3. Questioning the hypotheses 2038.4. Specific features of beautiful images detected by a computer 2048.4.1. Some observations on the photos in the AVA database 2058.4.2. The scores in the AVA database 207Conclusion 213Appendix 1. A Brief Review of Aesthetics 219Appendix 2. Aesthetics in China 237Appendix 3. The Aesthetic of Persian Miniatures 251Appendix 4. Aesthetics in Japan 263References 271Index 295

Henri Maître is Emeritus Professor at Télécom-Paris in France and was director of research at Télécom-Paris and the LTCI laboratory. He specializes in image processing and pattern recognition.

• Introduction: Image and Gaze xiChapter 1 The Legacy of Philosophers 11.1. The objectivist approach 31.1.1. The source: ancient Greece 31.1.2. After Greece 51.1.3. Kant and modern aesthetics 71.1.4. Objectivism after Kant: from pseudo-subjectivism to aesthetic realism 91.2. The subjectivist approach 131.2.1. From classicism to romanticism 141.2.2. The moderns 151.2.3. The influence of neurobiology 181.3. Subjectivism and objectivism: an ongoing debate 19Chapter 2. Neurobiology or the Arbitrator of Consciousness 252.1. fMRI protocols and neuroaesthetics 272.2. The fMRI quest for “beauty processes” in the brain 282.2.1. The role of the prefrontal cortex 282.2.2. The role of the insular cortex 312.2.3. The role of the visual areas 332.2.4. The role of memory and cognition 352.2.5. The role of embodiment 352.3. Responses from functional electric encephalography 362.4. A global cognitive scheme for aesthetic judgment? 392.4.1. J. Petitot’s neurogeometric model 402.4.2. A. Chatterjee’s aesthetic emotion model 402.4.3. The model by Brown et al. 422.4.4. Model proposed by H. Leder 432.4.5. The model by C. Redies 452.4.6. The emotions model developed by S. Koelsch et al. 472.4.7. L.H. Hsu’s model of emotions based on A. Damásio 472.4.8. Other models 502.5. A critique of neuroaesthetic methods 512.5.1. Criticism of neuroaesthetic methods 512.5.2. Criticisms of the objectives of neuroaesthetics 52Chapter 3. What Are the Criteria For a Beautiful Photo? 553.1. Before we enter into the fray 563.1.1. What reference books do we have? 563.1.2. “Beauty of an image” or “quality of an image”? 573.1.3. A glossary of aesthetic appraisal 583.1.4. Measuring beauty 603.2. Composition 633.2.1. Complexity versus simplicity 633.2.2. Unity 643.2.3. A specific case in composition: landscapes 643.2.4. Using oculometry to analyze composition 673.2.5. Format or aspect ratio 683.2.6. The rule of thirds (RoT) 703.2.7. The center of the image 723.2.8. Other rules for composition 733.3. Histograms, spectral properties and textures 763.3.1. Histograms and gray levels 763.3.2. Focus, spectral density, fractals 783.3.3. Textures 803.4. Color 823.4.1. About the concept of color 823.4.2. Preferences related to isolated colors 843.4.3. Preferences related to color palettes 863.5. What behavioral psychosociology has to say 933.5.1. Images of nature 933.5.2. The aesthetics of faces 963.5.3. The role of the signature, title and context 993.5.4. Perception and memory: prototypicality 101Chapter 4. Algorithmic Approaches to “Calculate” Beauty 1034.1. First steps: C. Henry 1034.2. G.D. Birkhoff’s mathematical approach 1044.3. Those who followed G.D. Birkhoff 1064.3.1. Beauty according to H.J. Eysenck 1064.3.2. The Post-War years: the designers, A. Moles and M. Bense 1064.3.3. A dynamic approach: P. Machado and A. Cardoso 1074.3.4. Work carried out by J. Rigau, M. Feixas and M. Bert 1084.4. Algorithmic approach with AI: J. Schmidhuber 110Chapter 5. The Holy Grail of the Digital World: Artificial Intelligence 1135.1. Which artificial intelligence? 1145.1.1. The principles 1145.1.2. Learning algorithms 1155.2. Why artificial intelligence in aesthetics? 1165.3. Expert opinions 1185.4. The database 1205.4.1. Generalist databases, used for aesthetic judgments 1225.4.2. Databases that are specialized for aesthetic photography 1265.4.3. Databases dedicated to artistic judgment 1295.4.4. Other image databases that are sometimes used 1305.4.5. Increasing databases 131Chapter 6. Primitive-based Classification Methods 1336.1. Judging aesthetics 1366.1.1. Multimedia primitives: the ACQUINE system (Datta et al.) 1366.1.2. Edges and chromatic distance: Ke et al. 1376.1.3. Photography rules: Luo and Tang and Mavridaki and Mezaris 1406.1.4. High-level primitives: Dhar et al. 1436.1.5. Generic descriptors of vision: Marchesotti et al. 1446.2. Help in composing beautiful photos 1486.2.1. The library of aesthetic primitives developed by Su et al. 1486.2.2. The OSCAR system by Yao et al. 1486.2.3. Embedded systems: Lo et al. and Wang et al. 1506.3. Some specific research related to the evaluation of aesthetics using primitives 1516.3.1. Color harmony: Lu et al. 1516.3.2. Group photography: Wang et al. 1536.3.3. Social networks and crowdsourcing: Schifanella et al. 1536.3.4. Looking at comments: San Pedro et al. 154Chapter 7. Deep Neural Network Systems 1557.1. DNNs dedicated to aesthetic evaluation 1577.1.1. High and low resolutions: the RAPID system, Lu et al. 1577.1.2. The multi-path DMA-Net architecture: Lu et al. 1607.1.3. Adapting to the size of the image: Mai et al. 1607.1.4. Finding beauty on the Web: Redi et al. 1637.1.5. Siamese and GAN networks: Kong et al. and Deng et al. 1657.1.6. Paying attention to the image construction: A-Lamp 1677.2. Variants around the basic DNN architecture 1707.2.1. Comparing photos between themselves: Schwarz et al. 1707.2.2. Making use of knowledge of the subject: Kao et al. 1727.2.3. BDN: halfway between classification and DNN 1747.2.4. Using the distribution of the evaluations 1757.2.5. Extracting a “dramatic” image from a panorama: the Creatism system 1787.3. Written appraisals: analyzing them and formulating new ones 1797.3.1. Photo critique captioning dataset (PCCD) 1817.3.2. Neural aesthetic image retriever (NAIR) 1827.3.3. Semantic processing by Ghosal et al. 1827.3.4. Aesthetic multi attribute network (AMAN) 1837.4. Measuring subjective beauty 1857.4.1. Recommendation systems 1867.4.2. Defining the user’s psychological profile 1887.4.3. Learning the user’s tastes through tests 1917.4.4. Multiplying concurrent expertise 194Chapter 8. A Critical Analysis of Machine Learning Techniques 1978.1. The popularity of studies on aesthetics 1978.2. A summary of learning methods 1998.2.1. Which architecture? Which software? 1998.2.2. What performances? 2008.3. Questioning the hypotheses 2038.4. Specific features of beautiful images detected by a computer 2048.4.1. Some observations on the photos in the AVA database 2058.4.2. The scores in the AVA database 207Conclusion 213Appendix 1. A Brief Review of Aesthetics 219Appendix 2. Aesthetics in China 237Appendix 3. The Aesthetic of Persian Miniatures 251Appendix 4. Aesthetics in Japan 263References 271Index 295