Visual DNA: representing and comparing images using distributions of neuron activations
<p>Selecting appropriate datasets is critical in modern computer vision. However, no general-purpose tools exist to evaluate the extent to which two datasets differ.</p> <p>For this, we propose representing images – and by extension datasets – using Distributions of Neuron Activat...
| Main Authors: | , , , |
|---|---|
| Format: | Conference item |
| Language: | English |
| Published: |
IEEE
2023
|
| _version_ | 1797110761283125248 |
|---|---|
| author | Ramtoula, B Gadd, M Newman, P De Martini, D |
| author_facet | Ramtoula, B Gadd, M Newman, P De Martini, D |
| author_sort | Ramtoula, B |
| collection | OXFORD |
| description | <p>Selecting appropriate datasets is critical in modern computer vision. However, no general-purpose tools exist to evaluate the extent to which two datasets differ.</p>
<p>For this, we propose representing images – and by extension datasets – using Distributions of Neuron Activations (DNAs). DNAs fit distributions, such as histograms or Gaussians, to activations of neurons in a pre-trained feature extractor through which we pass the image(s) to represent. This extractor is frozen for all datasets, and we rely on its generally expressive power in feature space. By comparing two DNAs, we can evaluate the extent to which two datasets differ with granular control over the comparison attributes of interest, providing the ability to customise the way distances are measured to suit the requirements of the task at hand. Furthermore, DNAs are compact, representing datasets of any size with less than 15 megabytes.</p>
<p>We demonstrate the value of DNAs by evaluating their applicability on several tasks, including conditional dataset comparison, synthetic image evaluation, and transfer learning, and across diverse datasets, ranging from synthetic cat images to celebrity faces and urban driving scenes.</p> |
| first_indexed | 2024-03-07T07:59:24Z |
| format | Conference item |
| id | oxford-uuid:b07d7d72-554a-410b-8979-00284688e102 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T07:59:24Z |
| publishDate | 2023 |
| publisher | IEEE |
| record_format | dspace |
| spelling | oxford-uuid:b07d7d72-554a-410b-8979-00284688e1022023-09-18T10:57:57ZVisual DNA: representing and comparing images using distributions of neuron activationsConference itemhttp://purl.org/coar/resource_type/c_5794uuid:b07d7d72-554a-410b-8979-00284688e102EnglishSymplectic ElementsIEEE2023Ramtoula, BGadd, MNewman, PDe Martini, D<p>Selecting appropriate datasets is critical in modern computer vision. However, no general-purpose tools exist to evaluate the extent to which two datasets differ.</p> <p>For this, we propose representing images – and by extension datasets – using Distributions of Neuron Activations (DNAs). DNAs fit distributions, such as histograms or Gaussians, to activations of neurons in a pre-trained feature extractor through which we pass the image(s) to represent. This extractor is frozen for all datasets, and we rely on its generally expressive power in feature space. By comparing two DNAs, we can evaluate the extent to which two datasets differ with granular control over the comparison attributes of interest, providing the ability to customise the way distances are measured to suit the requirements of the task at hand. Furthermore, DNAs are compact, representing datasets of any size with less than 15 megabytes.</p> <p>We demonstrate the value of DNAs by evaluating their applicability on several tasks, including conditional dataset comparison, synthetic image evaluation, and transfer learning, and across diverse datasets, ranging from synthetic cat images to celebrity faces and urban driving scenes.</p> |
| spellingShingle | Ramtoula, B Gadd, M Newman, P De Martini, D Visual DNA: representing and comparing images using distributions of neuron activations |
| title | Visual DNA: representing and comparing images using distributions of neuron activations |
| title_full | Visual DNA: representing and comparing images using distributions of neuron activations |
| title_fullStr | Visual DNA: representing and comparing images using distributions of neuron activations |
| title_full_unstemmed | Visual DNA: representing and comparing images using distributions of neuron activations |
| title_short | Visual DNA: representing and comparing images using distributions of neuron activations |
| title_sort | visual dna representing and comparing images using distributions of neuron activations |
| work_keys_str_mv | AT ramtoulab visualdnarepresentingandcomparingimagesusingdistributionsofneuronactivations AT gaddm visualdnarepresentingandcomparingimagesusingdistributionsofneuronactivations AT newmanp visualdnarepresentingandcomparingimagesusingdistributionsofneuronactivations AT demartinid visualdnarepresentingandcomparingimagesusingdistributionsofneuronactivations |