Skip to content

Commit

Permalink
More changes
Browse files Browse the repository at this point in the history
  • Loading branch information
@koutilya-pnvr
koutilya-pnvr committed Oct 4, 2023
1 parent 49f4d5a commit 51ca3bc
Showing 1 changed file with 23 additions and 13 deletions.
36 changes: 23 additions & 13 deletions index.html
View file
Original file line number Diff line number Diff line change
Expand Up @@ -118,7 +118,7 @@ <h3 class="title is-4">&#10024&#10024 Oral in ICCV 2023 &#10024&#10024</h3>
<a href="./assets/12703_poster.pdf"
class="external-link button is-normal is-rounded is-dark" target="_blank" rel="noopener noreferrer">
<span class="icon">
<i class="fas fa-file-pdf"></i>
<i class="fas fa-map"></i>
</span>
<span>Poster</span>
</a>
Expand All @@ -141,6 +141,15 @@ <h3 class="title is-4">&#10024&#10024 Oral in ICCV 2023 &#10024&#10024</h3>
</span>
<span>Video</span>
</a>
</span>
<span class="link-block">
<a href="https://prezi.com/view/9n4as5i8FMjAREBvnaPb/"
class="external-link button is-normal is-rounded is-dark" target="_blank" rel="noopener noreferrer">
<span class="icon">
<i class="fab fa-slideshare"></i>
</span>
<span>Prezi Slides</span>
</a>
</span>
<!-- Code Link. -->
<span class="link-block">
Expand Down Expand Up @@ -247,7 +256,7 @@ <h3 class="title is-4">Object-level semantics inside LDM</h3>
<figure class="is-centered">
<img src="./assets/Motivation_new1.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="Semantic information present in a pretrained LDM in coarsely segmenting novel categories"
width="50%" height="auto"/>
<figcaption>Coarse segmentation results from an LDM for two distinct images, demonstrating the encoding of fine-grained object-level semantic information within the model’s internal features.</figcaption>
</figure>
Expand All @@ -258,7 +267,7 @@ <h3 class="title is-4">Method</h3>
<figure class="is-centered">
<img src="./assets/LDZNet-Summary.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="Proposed LD-ZNet architecture"
width="100%" height="auto"/>
<figcaption>Overview of the proposed architecture</figcaption>
</figure>
Expand All @@ -277,7 +286,7 @@ <h3 class="title is-4">Phrasecut Dataset</h3>
<!-- <h2 class="title is-3">Visual Effects</h2>-->
<img src="./assets/Results_Phrasecut.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="Qualitative comparision of LD-ZNet with previous works on the PhraseCut dataset"
width="100%" height="auto"/>
</div>
</div>
Expand All @@ -290,7 +299,7 @@ <h3 class="title is-4">Phrasecut Dataset</h3>
<div class="column content has-text-centered">
<img src="./assets/Quantitative_Real.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="Quantitative comparision of LD-ZNet with previous works on the PhraseCut dataset"
width="70%" height="auto"/>
</div>

Expand All @@ -307,7 +316,7 @@ <h3 class="title is-4">AI-Generated Images (AIGI) Dataset</h3>
<figure class="is-centered">
<img src="./assets/AIGI_dataset.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="Examples from the AIGI dataset along with annotated labels of different objects and corresponding categorical captions"
width="90%" height="auto"/>
<figcaption>Examples from the AIGI dataset with annotations. Images gathered from the <a href="https://lexica.art/" target="_blank" rel="noopener noreferrer">lexica.art</a> website</figcaption>
</figure>
Expand All @@ -321,7 +330,7 @@ <h3 class="title is-4">AI-Generated Images (AIGI) Dataset</h3>
<h2 class="title is-3">Results on AIGI</h2>
<img src="./assets/Results_AI.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="Qualitative comparision of LD-ZNet with previous works on the AIGI dataset"
width="100%" height="auto"/>
</div>
</div>
Expand All @@ -335,13 +344,13 @@ <h2 class="title is-3">Results on AIGI</h2>
<br><br><br><br>
<img src="./assets/Quantitative_AIGI.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="Quantitative comparision of LD-ZNet with previous works on the AIGI dataset"
width="70%" height="auto"/>

<br><br>
<img src="./assets/Results_AI2.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="LD-ZNet can segment various objects from AIGI"
width="90%" height="auto"/>
</div>

Expand All @@ -360,9 +369,9 @@ <h2 class="title is-3">More qualitative results of LD-ZNet</h2>
<!-- <h2 class="title is-3">More qualitative results of LD-ZNet</h2>-->
<img src="./assets/Results_use_of_LDM_features.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="LD-ZNet can segment objects from animations, illustrations and celebrity images"
width="100%" height="auto"/>
<p>More qualitative examples where RGBNet fails to localize {``Guitar", ``Panda"} from animation images (top two rows), famous celebrities {``Scarlett Johansson", ``Kate Middleton"} (middle two rows) and objects such as {``Lamp", ``Trees"} from illustrations (bottom two rows). LD-ZNet benefits from using z combined with the internal LDM features to correctly segment these text prompts.</p>
<p>More qualitative examples where RGBNet fails to localize {``Guitar", ``Panda"} from animation images (top two rows),&nbsp; &nbsp; objects such as {``Lamp", ``Trees"} from illustrations&nbsp;(middle two rows) and famous celebrities {``Scarlett Johansson", ``Kate Middleton"} (bottom two rows). LD-ZNet benefits from using z combined with the internal LDM features to correctly segment these text prompts.</p>
</figure>
</div>
</div>
Expand All @@ -377,13 +386,14 @@ <h2 class="title is-3">More qualitative results of LD-ZNet</h2>
<h3 class="title is-4">Multi-object Segmentation</h3>
<img src="./assets/Results_Multi_segmentation.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="LD-ZNet has good understanding of the overall scene"
width="100%" height="auto"/>
<br><br><br>
<img src="./assets/Results_Multi_segmentation2.jpg"
class="center"
alt="LD-ZNet can segment various objects on real and AI-generated images"
alt="LD-ZNet has good understanding of the overall scene"
width="100%" height="auto"/>
<p>Multi-object segmentation on real and illustration images for various thing and stuff classes suggests LD-ZNet has a good understanding of the overall scene.</p>
</div>

</div>
Expand Down

0 comments on commit 51ca3bc

Please sign in to comment.