Google DeepMind's TIPSv2: AI That Truly Understands Images, Not Just Glances
AI image understanding currently has a core limitation.
When asked "What is in this picture?" it can provide a detailed response. However, asking "Where is the panda's left hind leg?" leads to vague answers. This is not a flaw in any specific model but a persistent issue across the entire visual-language large model domain: strong global understanding but weak local localization.
Google DeepMind introduced TIPSv2 in their latest paper, specifically designed to address this challenging problem.

The research team observed a counterintuitive finding: in fine-grained segmentation tasks, smaller student models frequently outperform larger teacher models. This happens because distillation removes the masking mechanism, compelling the model to learn every detail of the entire image, creating a form of "full-area supervision." Motivated by this insight, TIPSv2 introduced three key enhancements.
First, iBOT++. Traditional pre-training only computes loss for masked regions, leaving visible areas in a neglected state that causes local semantics to drift. iBOT++ requires the model to provide precise supervision over all visible areas, effectively upgrading the task from a "puzzle game" to "carefully reading the entire text." This single improvement boosted zero-shot segmentation performance by 14.1 percentage points.
Second, Head-only EMA. Traditional self-supervised training requires keeping two nearly identical large models in memory, which is highly resource-intensive. TIPSv2 discovered that the image-text contrastive loss alone is enough to stabilize the backbone network, so EMA only needs to be applied to the final projection head, eliminating the need to duplicate the backbone. This reduces the training parameter count by about 42%, making it faster with almost no performance drop.
Third, multi-granularity text pairing. During training, short web descriptions, medium-detail descriptions, and long descriptions generated by Gemini are randomly mixed and fed into the model, alternating between easy and hard tasks. This prevents the model from coasting on simple tasks while ensuring no details are overlooked.
The final results are compelling. TIPSv2 underwent frozen evaluation across nine tasks and 20 authoritative datasets. Zero-shot semantic segmentation achieved a new industry benchmark, while image-text retrieval and classification outperformed comparison models with 56% more parameters. Pure visual tasks also placed among the top performers.
The code and model weights for TIPSv2 are fully open-sourced. For teams working in medical imaging, autonomous driving, industrial inspection, and other domains that demand high-precision image understanding, this solution is well worth a close look.
Paper: https://www.alphaxiv.org/abs/2604.12012
Related article
Suno to Watermark Songs Amid Legal Battles
Suno, the platform enabling users to generate AI-created music, has unveiled new features to label platform-produced tracks, restrict downloads, and update community standards to curb unauthorized replicas. These updates arrive as Suno confronts mult
Musk Admits Grok Build Leaked User Code, Promises to Erase All Historical Data
Elon Musk directly addressed the privacy controversy surrounding Grok Build, beginning with a simple "True" to confirm the incident's validity. He pledged that all user data previously uploaded to SpaceXAI would be permanently erased, stating, "not a
U.S. Stocks Hit Historic Milestone as AI and Aerospace Giants Prepare for Trillion-Dollar Debut
Elon Musk, Sam Altman, and Dario Amodei, three titans of the technology sector, are advancing toward initial public offerings for their respective ventures. With SpaceX, OpenAI, and Anthropic—three industry behemoths nearing trillion-dollar valuation
Related Special Topic Recommendations
Comments (0)
0/500
AI image understanding currently has a core limitation.
When asked "What is in this picture?" it can provide a detailed response. However, asking "Where is the panda's left hind leg?" leads to vague answers. This is not a flaw in any specific model but a persistent issue across the entire visual-language large model domain: strong global understanding but weak local localization.
Google DeepMind introduced TIPSv2 in their latest paper, specifically designed to address this challenging problem.

The research team observed a counterintuitive finding: in fine-grained segmentation tasks, smaller student models frequently outperform larger teacher models. This happens because distillation removes the masking mechanism, compelling the model to learn every detail of the entire image, creating a form of "full-area supervision." Motivated by this insight, TIPSv2 introduced three key enhancements.
First, iBOT++. Traditional pre-training only computes loss for masked regions, leaving visible areas in a neglected state that causes local semantics to drift. iBOT++ requires the model to provide precise supervision over all visible areas, effectively upgrading the task from a "puzzle game" to "carefully reading the entire text." This single improvement boosted zero-shot segmentation performance by 14.1 percentage points.
Second, Head-only EMA. Traditional self-supervised training requires keeping two nearly identical large models in memory, which is highly resource-intensive. TIPSv2 discovered that the image-text contrastive loss alone is enough to stabilize the backbone network, so EMA only needs to be applied to the final projection head, eliminating the need to duplicate the backbone. This reduces the training parameter count by about 42%, making it faster with almost no performance drop.
Third, multi-granularity text pairing. During training, short web descriptions, medium-detail descriptions, and long descriptions generated by Gemini are randomly mixed and fed into the model, alternating between easy and hard tasks. This prevents the model from coasting on simple tasks while ensuring no details are overlooked.
The final results are compelling. TIPSv2 underwent frozen evaluation across nine tasks and 20 authoritative datasets. Zero-shot semantic segmentation achieved a new industry benchmark, while image-text retrieval and classification outperformed comparison models with 56% more parameters. Pure visual tasks also placed among the top performers.
The code and model weights for TIPSv2 are fully open-sourced. For teams working in medical imaging, autonomous driving, industrial inspection, and other domains that demand high-precision image understanding, this solution is well worth a close look.
Paper: https://www.alphaxiv.org/abs/2604.12012
Suno to Watermark Songs Amid Legal Battles
Suno, the platform enabling users to generate AI-created music, has unveiled new features to label platform-produced tracks, restrict downloads, and update community standards to curb unauthorized replicas. These updates arrive as Suno confronts mult
Musk Admits Grok Build Leaked User Code, Promises to Erase All Historical Data
Elon Musk directly addressed the privacy controversy surrounding Grok Build, beginning with a simple "True" to confirm the incident's validity. He pledged that all user data previously uploaded to SpaceXAI would be permanently erased, stating, "not a
U.S. Stocks Hit Historic Milestone as AI and Aerospace Giants Prepare for Trillion-Dollar Debut
Elon Musk, Sam Altman, and Dario Amodei, three titans of the technology sector, are advancing toward initial public offerings for their respective ventures. With SpaceX, OpenAI, and Anthropic—three industry behemoths nearing trillion-dollar valuation





Home






