PageSourceSearch

https://khawar-islam.github.io/

html khawar-islam.github.io collected 2026-10-03 09:27:24 UTC 21,477 bytes, 498 lines download raw bytes

1<!doctype html>
2<html lang="en">
3<head>
4  <meta charset="utf-8">
5  <title>Khawar Islam</title>
6  <meta name="viewport" content="width=device-width, initial-scale=1">
7  <meta name="description" content="Khawar Islam — PhD student in Computer Science at the University of Melbourne. Research on data augmentation, diffusion models, video understanding, and robust computer vision.">
8  <style>
9    * { box-sizing: border-box; }
10    body {
11      margin: 0;
12      background: #fff;
13      color: #333;
14      font: 15px/1.6 -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif;
15    }
16    a { color: #1a73a7; text-decoration: none; }
17    a:hover { color: #0d4f75; text-decoration: underline; }
18    .container {
19      max-width: 1080px;
20      margin: 0 auto;
21      padding: 2.5rem 1.5rem;
22      display: flex;
23      gap: 3rem;
24      align-items: flex-start;
25    }
26    /* ---------- Sidebar ---------- */
27    .sidebar {
28      width: 240px;
29      flex-shrink: 0;
30      position: sticky;
31      top: 2.5rem;
32      text-align: center;
33    }
34    .sidebar img.profile {
35      width: 200px;
36      border-radius: 6px;
37      margin-bottom: 1rem;
38    }
39    .sidebar h1 {
40      font-size: 1.3rem;
41      font-weight: 600;
42      color: #222;
43      margin: 0 0 0.75rem;
44    }
45    .sidebar p { margin: 0 0 0.75rem; font-size: 0.88rem; color: #555; }
46    .sidebar .role { color: #333; }
47    .sidebar .links { font-size: 0.88rem; line-height: 2; }
48    /* ---------- Main ---------- */
49    .main { flex: 1; min-width: 0; }
50    .main h2 {
51      font-size: 1.25rem;
52      font-weight: 600;
53      color: #222;
54      margin: 2.2rem 0 0.9rem;
55      padding-bottom: 0.35rem;
56      border-bottom: 1px solid #e5e5e5;
57      scroll-margin-top: 3.2rem;
58    }
59    .main h2:first-child { margin-top: 0; }
60    /* In-page section nav */
61    .site-nav {
62      display: flex;
63      flex-wrap: wrap;
64      gap: 0.15rem 1.35rem;
65      margin: 0 0 1.6rem;
66      padding: 0.65rem 0;
67      border-top: 1px solid #e5e5e5;
68      border-bottom: 1px solid #e5e5e5;
69      position: sticky;
70      top: 0;
71      background: #fff;
72      z-index: 10;
73    }
74    .site-nav a {
75      color: #444;
76      font-weight: 600;
77      font-size: 0.92rem;
78      text-decoration: none;
79    }
80    .site-nav a:hover { color: #1a73a7; text-decoration: underline; }
81    .venue { color: #b03030; font-weight: 600; font-style: italic; }
82    /* News */
83    ul.news { list-style: none; padding: 0; margin: 0; }
84    ul.news li { margin-bottom: 0.45rem; padding-left: 5.2em; text-indent: -5.2em; }
85    ul.news .date { display: inline-block; width: 5.2em; text-indent: 0; color: #888; font-variant-numeric: tabular-nums; }
86    /* Service appointments */
87    ul.service { list-style: none; padding: 0; margin: 0; }
88    ul.service li { margin-bottom: 0.55rem; }
89    .badge {
90      display: inline-block;
91      padding: 0.12em 0.5em;
92      margin: 0.1em 0.15em 0.1em 0;
93      border-radius: 3px;
94      font-size: 0.82rem;
95      font-weight: 600;
96      line-height: 1.5;
97      color: #fff;
98      vertical-align: middle;
99    }
100    .badge-reviewer { background: #c9a227; }
101    .badge-journal { background: #2e7d4f; }
102    .badge-volunteer { background: #b03030; }
103    /* Publications — text-only list (Sanghyuk Chun style) */
104    ul.pubs {
105      list-style: none;
106      padding: 0;
107      margin: 0 0 0.5rem;
108    }
109    ul.pubs > li {
110      margin-bottom: 1.15rem;
111      padding-bottom: 1.05rem;
112      border-bottom: 1px solid #f0f0f0;
113    }
114    ul.pubs > li:last-child { border-bottom: none; padding-bottom: 0; }
115    ul.pubs .title { font-weight: 600; color: #222; }
116    ul.pubs .meta {
117      list-style: none;
118      padding: 0.2rem 0 0;
119      margin: 0;
120    }
121    ul.pubs .meta li { margin-bottom: 0.1rem; }
122    ul.pubs .authors b { font-weight: 600; }
123    ul.pubs .plinks { font-size: 0.92rem; }
124    /* See more (details/summary, same pattern as Sanghyuk Chun) */
125    details.see-more { margin: 0.6rem 0 1rem; }
126    details.see-more > summary {
127      cursor: pointer;
128      color: #1a73a7;
129      font-weight: 600;
130      list-style: none;
131      user-select: none;
132    }
133    details.see-more > summary::-webkit-details-marker { display: none; }
134    details.see-more > summary::before { content: "▸ "; }
135    details.see-more[open] > summary::before { content: "▾ "; }
136    details.see-more > summary:hover { color: #0d4f75; text-decoration: underline; }
137    details.see-more ul.news,
138    details.see-more ul.pubs { margin-top: 0.75rem; }
139    /* Generic lists */
140    .main ul { padding-left: 1.2rem; margin: 0 0 1rem; }
141    .main ul li { margin-bottom: 0.3rem; }
142    .main h3 {
143      font-size: 1rem;
144      font-weight: 600;
145      color: #333;
146      margin: 1.2rem 0 0.5rem;
147    }
148    footer {
149      font-size: 0.8rem;
150      color: #999;
151      margin-top: 3rem;
152      border-top: 1px solid #eee;
153      padding-top: 1rem;
154    }
155    /* ---------- Responsive ---------- */
156    @media (max-width: 760px) {
157      .container { flex-direction: column; gap: 1.5rem; }
158      .sidebar { position: static; width: 100%; }
159      .site-nav { gap: 0.35rem 1rem; }
160    }
161  </style>
162</head>
163<body>
164<div class="container">
165
166  <!-- ================= Sidebar ================= -->
167  <aside class="sidebar">
168    <img class="profile" src="assets/img/khawarislam.jpeg" alt="Khawar Islam">
169    <h1>Khawar Islam</h1>
170    <p class="role">
171      Ph.D. Student (2024&ndash;Present)<br>
172      University of Melbourne<br>
173      Melbourne, Australia
174    </p>
175    <p class="role">
176      Senior Scientist &ndash; Autonomy AI<br>
177      <a href="https://www.neubility.ai/">Neubility</a>, Seoul, South Korea
178    </p>
179    <p>khawarr [dot] islam [at] gmail [dot] com</p>
180    <p class="links">
181      <a href="assets/cv.pdf">CV</a> |
182      <a href="https://scholar.google.com.pk/citations?user=SFClxNEAAAAJ&amp;hl=en">Google Scholar</a> |
183      <a href="https://dblp.org/pid/261/2710.html">dblp</a><br>
184      <a href="https://github.com/khawar-islam">GitHub</a> |
185      <a href="https://www.linkedin.com/in/khawar-islam/">LinkedIn</a>
186    </p>
187  </aside>
188
189  <!-- ================= Main ================= -->
190  <main class="main">
191
192    <nav class="site-nav" aria-label="Page sections">
193      <a href="#news">News</a>
194      <a href="#publications">Publications</a>
195      <a href="#patents">Patents</a>
196      <a href="#projects">Projects</a>
197      <a href="#awards">Awards</a>
198      <a href="#activities">Activities</a>
199      <a href="#skills">Skills</a>
200    </nav>
201
202    <h2 id="about">About Me</h2>
203    <p>
204      I am a Ph.D. student in Computer Science and Engineering at the University of Melbourne,
205      advised by Prof. <a href="https://scholar.google.com/citations?user=xE_0DiwAAAAJ">Naveed Akhtar</a>.
206      My thesis focuses on benchmarking video understanding. My broader research interests include
207      the generalization and robustness of neural networks, responsible data augmentation with
208      diffusion models, vision-language models, and consistent character generation.
209    </p>
210    <p>
211      Alongside my PhD, I work as a Senior Scientist &ndash; Autonomy AI at
212      <a href="https://www.neubility.ai/">Neubility</a>. Previously, I was Chief Technology Officer
213      (CTO) &ndash; Gen AI at <a href="https://upendi.app/">Upendi.app</a>
213, a Research Scientist &ndash;
214      CV/ML at <a href="http://www.coremaxtech.com/">CoreMax Technology</a>, and a Research Associate
215      at Sejong University, all in Seoul, South Korea.
216    </p>
217    <p>
218      I received my M.S. in Computer Science and Engineering from Sejong University, Seoul, where I
219      worked on image understanding and reconstruction, including an image compression method
220      presented at CVPRw. I have attended several summer schools and served as a technical volunteer
221      at top AI/ML conferences.
222    </p>
223
224    <h2 id="news">News</h2>
225    <ul class="news">
226      <li><span class="date">07/2026</span> Two papers (S&sup2;-FracMix, HSFM) accepted at <b>ECCV 2026</b>!</li>
227      <li><span class="date">06/2026</span> <b>US Patent 12,657,783 B1</b> granted.</li>
228      <li><span class="date">04/2026</span> GenMix accepted in <b>Expert Systems with Applications</b> (Elsevier).</li>
229      <li><span class="date">03/2025</span> One paper on responsible data augmentation accepted at the <b>ICLR 2025</b> workshop track.</li>
230      <li><span class="date">09/2024</span> Starting my Ph.D. at the <b>University of Melbourne</b>, Australia.</li>
231    </ul>
232    <details class="see-more">
233      <summary>See older news</summary>
234      <ul class="news">
235        <li><span class="date">02/2024</span> DiffuseMix accepted at <b>CVPR 2024</b>!</li>
236        <li><span class="date">10/2022</span> Face Pyramid Vision Transformer accepted at <b>BMVC 2022</b>.</li>
237      </ul>
238    </details>
239
240    <h2 id="current_service">Current Service Appointments</h2>
241    <ul class="service">
242      <li>
243        <strong>Reviewer</strong> &ndash;
244        <span class="badge badge-journal">TPAMI</span>
245        <span class="badge badge-journal">IJCV</span>
246        <span class="badge badge-journal">CVIU</span>
247      </li>
248      <li>
249        <strong>Reviewer</strong> &ndash;
250        <span class="badge badge-reviewer">CVPR</span>
251        <span class="badge badge-reviewer">ICCV</span>
252        <span class="badge badge-reviewer">ECCV</span>
253        <span class="badge badge-reviewer">NeurIPS</span>
254        <span class="badge badge-reviewer">ICML</span>
255        <span class="badge badge-reviewer">ICLR</span>
256        <span class="badge badge-reviewer">AAAI</span>
257        <span class="badge badge-reviewer">WACV</span>
258      </li>
259      <li>
260        <strong>Volunteer</strong> &ndash;
261        <span class="badge badge-volunteer">NeurIPS</span>
262        <span class="badge badge-volunteer">ICML</span>
263        <span class="badge badge-volunteer">ICLR</span>
264        <span class="badge badge-volunteer">IJCAI</span>
265        <span class="badge badge-volunteer">AAAI</span>
266        <span class="badge badge-volunteer">CVPR</span>
267      </li>
268    </ul>
269
270    <h2 id="publications">Publications</h2>
271    <ul class="pubs">
272      <li>
273        <span class="title">S&sup2;-FracMix: Label-Preserving Self-Saliency Mixup Augmentation</span>
274        <ul class="meta">
275          <li class="authors"><b>Khawar Islam</b>, Arif Mahmood, Xin Jin, Naveed Akhtar</li>
276          <li><span class="venue">ECCV 2026.</span> <span class="plinks">paper coming soon</span></li>
277        </ul>
278      </li>
279      <li>
280        <span class="title">HSFM: Hard-Set-Guided Feature-Space Meta-Learning for Robust Classification under Spurious Correlations</span>
281        <ul class="meta">
282          <li class="authors">Aryan Yazdan Parast, <b>Khawar Islam</b>, Soyoun Won, Basim Azam, Naveed Akhtar</li>
283          <li><span class="venue">ECCV 2026.</span> <span class="plinks">paper coming soon</span></li>
284        </ul>
285      </li>
286      <li>
287        <span class="title">GenMix: Effective Data Augmentation with Generative Diffusion Model Image Editing</span>
288        <ul class="meta">
289          <li class="authors"><b>Khawar Islam</b>, Muhammad Zaigham Zaheer, Arif Mahmood, Karthik Nandakumar, Naveed Akhtar</li>
290          <li>
291            <span class="venue">Expert Systems with Applications (ESWA), 2026.</span>
292            <span class="plinks">
293              <a href="https://www.sciencedirect.com/science/article/pii/S0957417426011863">paper</a> |
294              <a href="https://arxiv.org/abs/2412.02366">arXiv</a> |
295              <a href="https://github.com/khawar-islam/GenMix">
295code</a>
296            </span>
297          </li>
298        </ul>
299      </li>
300      <li>
301        <span class="title">Context-Guided Responsible Data Augmentation with Diffusion Models</span>
302        <ul class="meta">
303          <li class="authors"><b>Khawar Islam</b>, Naveed Akhtar</li>
304          <li><span class="venue">ICLR 2025 Workshop.</span> <span class="plinks"><a href="https://arxiv.org/abs/2503.10687">paper</a> | <a href="https://github.com/khawar-islam/DiffCoRe-Mix">code</a></span></li>
305        </ul>
306      </li>
307      <li>
308        <span class="title">DiffuseMix: Label-Preserving Data Augmentation with Diffusion Models</span>
309        <ul class="meta">
310          <li class="authors"><b>Khawar Islam</b>, Muhammad Zaigham Zaheer, Arif Mahmood, Karthik Nandakumar</li>
311          <li>
312            <span class="venue">CVPR 2024.</span>
313            <span class="plinks">
314              <a href="https://arxiv.org/abs/2405.14881">paper</a> |
315              <a href="https://diffusemix.github.io/">project page</a> |
316              <a href="https://github.com/khawar-islam/diffuseMix">code</a>
317            </span>
318          </li>
319        </ul>
320      </li>
321    </ul>
322
323    <details class="see-more">
324      <summary>See more publications</summary>
325      <ul class="pubs">
326        <li>
327          <span class="title">Face Pyramid Vision Transformer</span>
328          <ul class="meta">
329            <li class="authors"><b>Khawar Islam</b>, Muhammad Zaigham Zaheer, Arif Mahmood</li>
330            <li>
331              <span class="venue">BMVC 2022.</span>
332              <span class="plinks">
333                <a href="https://khawar-islam.github.io/fpvt/">project page</a> |
334                <a href="https://bmvc2022.mpi-inf.mpg.de/0758_supp.pdf">supplementary</a> |
335                <a href="https://github.com/khawar-islam/FPVT_BMVC22">code</a> |
336                <a href="https://bmvc2022.mpi-inf.mpg.de/0758_poster.pdf">poster</a> |
337                <a href="https://bmvc2022.mpi-inf.mpg.de/0758_video.mp4">video</a>
338              </span>
339            </li>
340          </ul>
341        </li>
342        <li>
343          <span class="title">Image Compression with Recurrent Neural Network and Generalized Divisive Normalization</span>
344          <ul class="meta">
345            <li class="authors"><b>Khawar Islam</b>, Dang Lien Minh, Sujin Lee, Hyeonjoon Moon</li>
346            <li>
347              <span class="venue">CVPR 2021 Workshop.</span>
348              <span class="plinks">
349                <a href="https://openaccess.thecvf.com/content/CVPR2021W/CLIC/papers/Islam_Image_Compression_With_Recurrent_Neural_Network_and_Generalized_Divisive_Normalization_CVPRW_2021_paper.pdf">paper</a> |
350                <a href="https://khawar-islam.github.io/cvpr/">project page</a> |
351                <a href="https://github.com/khawar-islam/ImageCompression">code</a> |
352                <a href="https://ieeexplore.ieee.org/document/9522707">IEEE Xplore</a>
353              </span>
354            </li>
355          </ul>
356        </li>
357        <li>
358          <span class="title">Face Recognition Using Shallow Age-Invariant Data</span>
359          <ul class="meta">
360            <li class="authors"><b>Khawar Islam</b>, Sujin Lee, Dongil Han, Hyeonjoon Moon</li>
361            <li>
362              <span class="venue">IVCNZ 2021.</span>
363              <span class="plinks"><a href="https://ieeexplore.ieee.org/document/9653432">IEEE Xplore</a></span>
364            </li>
365          </ul>
366        </li>
367        <li>
368          <span class="title">Person Search: New Paradigm of Person Re-Identification: A Survey and Outlook of Recent Works</span>
369          <ul class="meta">
370            <li class="authors"><b>Khawar Islam</b></li>
371            <li>
372              <span class="venue">Image and Vision Computing (IMAVIS), 2020.</span>
373              <span class="plinks"><a href="https://www.sciencedirect.com/science/article/pii/S0262885620301025">paper</a></span>
374            </li>
375          </ul>
376        </li>
377      </ul>
378    </details>
379
380    <h2 id="patents">Patents</h2>
381    <ul class="pubs">
382      <li>
383        <span class="title">System and Method for Label-Preserving Data Augmentation with Diffusion Models</span>
384        <ul class="meta">
385          <li class="authors"><b>Khawar Islam</b>, Muhammad Zaigham Zaheer, Arif Mahmood, Karthik Nandakumar</li>
386          <li>
387            <span class="venue">US Patent 12,657,783 B1, granted June 16, 2026.</span>
388            <span class="plinks">
389              <a href="https://patents.google.com/patent/US12657783B1">Google Patents</a> |
390              <a href="assets/US-12657783-B1.pdf">PDF</a>
391            </span>
392          </li>
393        </ul>
394      </li>
395    </ul>
396
397    <h2 id="awards">Awards &amp; Honors</h2>
398    <ul>
399      <li>NeurIPS Support, Sponsored by Naver Labs Europe, DeepMind, and Google AI, 2021</li>
400      <li>Full Tuition Fee Waiver, School of Computer Engineering, Sejong University, 2020</li>
401      <li>Prof. Stipend, CVPR Lab, Sejong University, 2020</li>
402      <li>Conference Travel Grant, Ministry of Planning Commission, Pakistan, 2020</li>
403      <li>Merit Scholarship for Bachelor Degree, Quaid-e-Azam Aligarh Society, Karachi, 2018</li>
404      <li>Most Valuable Professional Award, C-SharpCorner, India, 2016 &amp; 2017</li>
405      <li>Laptop Award, Prime Minister Pakistan Program, Karachi, Pakistan, 2016</li>
406    </ul>
407
408    <h2 id="activities">Academic Activities</h2>
409
410    <h3>Academic Services</h3>
411    <ul>
412      <li>Technical Volunteer &ndash; NeurIPS 2021, IJCAI 2021, ICML 2021, ICLR 2021, AAAI 2021</li>
413      <li>Student Volunteer &ndash; CVPR 2021</li>
414    </ul>
415
416    <h3>Reviewer &ndash; International Journals</h3>
417    <ul>
418      <li>IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)</li>
419      <li>International Journal of Computer Vision (IJCV)</li>
420      <li>Computer Vision and Image Understanding (CVIU)</li>
421    </ul>
422
423    <h3>Reviewer &ndash; International Conferences</h3>
424    <ul>
425      <li>CVPR 2019&ndash;
4252023, ICCV 2019&ndash;2023, ECCV 2020&ndash;2022</li>
426      <li>NeurIPS 2020&ndash;2022, ICML 2020&ndash;2023, ICLR 2021&ndash;2023</li>
427      <li>AAAI 2020&ndash;2023, WACV 2020</li>
428    </ul>
429
430    <h3>Moderator</h3>
431    <ul>
432      <li>AI Deep Group on Facebook, 23.1K members</li>
433      <li>Artificial Intelligence, Machine and Deep Learning Group, 754.4K members</li>
434      <li>Deep Learning and Machine Learning, 109.5K members</li>
435      <li>Thailand Deep Learning, 19.8K members</li>
436      <li>Computer Vision, 129.6K members</li>
437      <li>Computer Vision and Machine Learning, 17.5K members</li>
438    </ul>
439
440    <h2 id="projects">Projects</h2>
441    <ul>
442      <li><b>High-Quality Video Generation from Diffusion Models (Text-2-Video), 2024&ndash;Present</b><br>
443        Designed identity-specific prompts and fine-tuned models for video generation.</li>
444      <li><b>Image Restoration and Enhancement with GANs and Diffusion Models, 2024&ndash;Present</b><br>
445        Worked on image super-resolution, deblurring, in-painting, and dehazing.</li>
446      <li><b>Image Editing Instructions with Diffusion Models (Image-to-Image), 2023</b><br>
447        Integrated Imagic, SINE, and LEDITS for high-resolution image editing.</li>
448      <li><b>Foundation Multimodal Vision Language Models, 2023</b><br>
449        Generated captions and evaluated models on various metrics.</li>
450      <li><b>Vision Language Models for Unseen Domains &amp; Domain Shift, 2023</b><br>
451        Fine-tuned models to improve robustness and performed experiments on multiple datasets.</li>
452      <li><b>Reliable Lightweight Real-Time Open-Vocabulary Object Detection, 2023</b><br>
453        Investigated and fine-tuned models for state-of-the-art performance.</li>
454      <li><b>Online Continual Learning with Blurry Data and Incorrect Labels, 2022</b><br>
455        Developed a framework for diversity and purity in memory updates.</li>
456      <li><b>Vision Transformer for General and Age-Invariant Face Recognition, 2021</b><br>
457        Proposed a new ViT architecture and conducted experiments on face aging datasets.</li>
458    </ul>
459
460    <h2 id="skills">Technical Skills</h2>
461    <ul>
462      <li><b>IP Algorithms:</b> Segmentation, Filter Design, Noise Removal, Compression, Super Resolution</li>
463      <li><b>Object Detection:</b> SSD, YOLO-NAS, YOLOX, Faster R-CNN, DETR, GroundingDINO, YOLOv8</li>
464      <li><b>DL Techniques:</b> Tracking, Optical Flow, Particle Filtering, Pose Estimation</li>
465      <li><b>CV Algorithms:</b> Image Classification and Detection, Tracking, Siamese Networks</li>
466      <li><b>Software Control:</b> JIRA, Agile, Scrum, GIT, Subversion</li>
467      <li><b>Languages:</b> Python, Keras, JAVA, Unix Shell Scripting</li>
468      <li><b>Deployment:</b> Docker, Kubernetes</li>
469      <li><b>Editors:</b> Cursor AI, Windsurf, Visual Studio Code, PyCharm, Jupyter Notebook</li>
470      <li><b>Operating Systems:</b> Windows 10, Ubuntu 18.04 / 20.04, macOS</li>
471      <li><b>APIs:</b> PyTorch, TensorFlow, Caffe, OpenCV, Flask, Django</li>
472    </ul>
473
474    <h2 id="schools">Summer and Winter Schools</h2>
475    <ul>
476      <li>OxML School, Oxford Machine Learning Summer School, 2022</li>
477      <li>NYU AI School, New York Artificial Intelligence School, USA, 2022</li>
478      <li>MLSS, Machine Learning Summer School, Taipei, Taiwan, 2021</li>
479      <li>EEML, Eastern European Machine Learning Summer School, 2021</li>
480    </ul>
481
482    <h2 id="mentorship">Q/A Mentorship Sessions</h2>
483    <ul>
484      <li>Yingzhen Li, Mentor at ICML, Imperial College London</li>
485      <li>Evan Shelhamer, Mentor at ICML, DeepMind, Google</li>
486      <li>Wei-Lun (Harry) Chao, Mentor at ICLR, Ohio State University</li>
487      <li>Shakir Mohamed, Mentor at ICLR, DeepMind, Google</li>
488      <li>Emmanuel Kahembwe, Mentor at ICLR, VDE (UK &amp; Ireland)</li>
489    </ul>
490
491    <footer>
492      Layout inspired by <a href="https://sanghyukchun.github.io/home/">Sanghyuk Chun's homepage</a>.
493    </footer>
494
495  </main>
496</div>
497</body>
498</html>

Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.