Bolin Lai

Bolin Lai, Miao Liu, Fiona Ryan et al. · gatech

In this paper, we present the first transformer-based model to address the challenging problem of egocentric gaze estimation. We observe that the connection between the global scene context and local visual information is vital for localizing the gaze fixation from egocentric video frames. To this end, we design the transformer encoder to embed the global context as one additional visual token and further propose a novel Global-Local Correlation (GLC) module to explicitly model the correlation of the global token and each local token. We validate our model on two egocentric video datasets - EGTEA Gaze+ and Ego4D. Our detailed ablation studies demonstrate the benefits of our method. In addition, our approach exceeds previous state-of-the-arts by a large margin. We also provide additional visualizations to support our claim that global-local correlation serves a key representation for predicting gaze fixation from egocentric videos. More details can be found in our website (https://bolinlai.github.io/GLC-EgoGazeEst).

Bolin Lai, Hongxin Zhang, Miao Liu et al. · gatech

Persuasion modeling is a key building block for conversational agents. Existing works in this direction are limited to analyzing textual dialogue corpus. We argue that visual signals also play an important role in understanding human persuasive behaviors. In this paper, we introduce the first multimodal dataset for modeling persuasion behaviors. Our dataset includes 199 dialogue transcriptions and videos captured in a multi-player social deduction game setting, 26,647 utterance level annotations of persuasion strategy, and game level annotations of deduction game outcomes. We provide extensive experiments to show how dialogue context and visual signals benefit persuasion strategy prediction. We also explore the generalization ability of language models for persuasion modeling and the role of persuasion strategies in predicting social deduction game outcomes. Our dataset, code, and models can be found at https://persuasion-deductiongame.socialai-data.org.

Xu Cao, Pranav Virupaksha, Wenqi Jia et al. · gatech

Previous research in human gesture recognition has largely overlooked multi-person interactions, which are crucial for understanding the social context of naturally occurring gestures. This limitation in existing datasets presents a significant challenge in aligning human gestures with other modalities like language and speech. To address this issue, we introduce SocialGesture, the first large-scale dataset specifically designed for multi-person gesture analysis. SocialGesture features a diverse range of natural scenarios and supports multiple gesture analysis tasks, including video-based recognition and temporal localization, providing a valuable resource for advancing the study of gesture during complex social interactions. Furthermore, we propose a novel visual question answering (VQA) task to benchmark vision language models'(VLMs) performance on social gesture understanding. Our findings highlight several limitations of current gesture recognition models, offering insights into future directions for improvement in this field. SocialGesture is available at huggingface.co/datasets/IrohXu/SocialGesture.

Junho Kim, Xu Cao, Houze Yang et al.

Understanding social interactions requires reasoning over subtle non-verbal cues, yet current multimodal large language models (MLLMs) often fail to identify who interacts with whom in multi-person videos. We introduce GRASP, a large-scale social reasoning dataset that connects high-level social QA with fine-grained gaze and deictic gesture events. GRASP contains 290K question--answer pairs over 46K videos totaling 749 hours, organized by a 16-category taxonomy spanning gaze, gesture, and joint gaze--gesture reasoning, together with GRASP-Bench for evaluation. Unlike prior resources that focus on either isolated cues or high-level social QA, GRASP builds questions from identity-consistent gaze trajectories, deictic gestures, and their joint compositions into social events. Moreover, we propose Social Grounding Reward (SGR), a learning signal that uses these social events to encourage models to reason about the participants involved in each interaction. Experiments show that SGR improves performance on GRASP-Bench while maintaining zero-shot performance on related social video QA benchmarks.

Xinpeng Li, Shijian Deng, Bolin Lai et al.

Multimodal social interaction understanding (MMSI) is critical in human-robot interaction systems. In real-world scenarios, AI agents are required to provide real-time feedback. However, existing models often depend on both past and future contexts, which hinders them from applying to real-world problems. To bridge this gap, we propose an online MMSI setting, where the model must resolve MMSI tasks using only historical information, such as recorded dialogues and video streams. To address the challenges of missing the useful future context, we develop a novel framework, named Online-MMSI-VLM, that leverages two complementary strategies: multi-party conversation forecasting and social-aware visual prompting with multi-modal large language models. First, to enrich linguistic context, the multi-party conversation forecasting simulates potential future utterances in a coarse-to-fine manner, anticipating upcoming speaker turns and then generating fine-grained conversational details. Second, to effectively incorporate visual social cues like gaze and gesture, social-aware visual prompting highlights the social dynamics in video with bounding boxes and body keypoints for each person and frame. Extensive experiments on three tasks and two datasets demonstrate that our method achieves state-of-the-art performance and significantly outperforms baseline models, indicating its effectiveness on Online-MMSI. The code and pre-trained models will be publicly released at: https://github.com/Sampson-Lee/OnlineMMSI.

Wenqian Ye, Bohan Liu, Guangtao Zheng et al.

Spurious bias, a tendency to exploit spurious correlations between superficial input attributes and prediction targets, has revealed a severe robustness pitfall in classical machine learning problems. Multimodal Large Language Models (MLLMs), which leverage pretrained vision and language models, have recently demonstrated strong capability in joint vision-language understanding. However, both the presence and severity of spurious biases in MLLMs remain poorly understood. In this work, we address this gap by analyzing the spurious biases in the multimodal setting and uncovering the specific inference-time data patterns that can manifest this problem. To support this analysis, we introduce MM-SpuBench, a comprehensive, human-verified benchmark dataset consisting of image-class pairs annotated with core and spurious attributes, grounded in our taxonomy of nine distinct types of spurious correlations. The benchmark is constructed using human-interpretable attribute information to capture a wide range of spurious patterns reflective of real-world knowledge. Leveraging this benchmark, we conduct a comprehensive evaluation of the state-of-the-art open-source and proprietary MLLMs with both standard accuracy and the proposed Conditional Generation Likelihood Advantage (CGLA). Our findings highlight the persistence of reliance on spurious correlations and the difficulty of mitigation on our benchmark. We hope this work can inspire new technical strides to mitigate these biases. Our benchmark is publicly available at https://huggingface.co/datasets/mmbench/MM-SpuBench.

Xu Cao, Yifan Shen, Bolin Lai et al.

Recently, Multimodal Large Language Models (MLLMs) and Vision Language Models (VLMs) have shown great promise in language-guided perceptual tasks such as recognition, segmentation, and object detection. However, their effectiveness in addressing visual cognition problems that require high-level multi-image reasoning and visual working memory is not well-established. One such challenge is matrix reasoning - the cognitive ability to discern relationships among patterns in a set of images and extrapolate to predict subsequent patterns. This skill is crucial during the early neurodevelopmental stages of children. Inspired by the matrix reasoning tasks in Raven's Progressive Matrices (RPM) and Wechsler Intelligence Scale for Children (WISC), we propose a new dataset MaRs-VQA to evaluate the visual cognition capability of MLLMs and compare their performance with existing human visual cognition studies. Based on the training data of MaRs-VQA, we also finetune a baseline model Qwen2-VCog with multi-stage cognition reasoning annotations. Our comparative experiments with different baselines reveal a gap between MLLMs and human intelligence, highlighting the visual cognitive limitations of current MLLMs. We believe that the public release of MaRs-VQA and the Qwen2-VCog baseline model will drive progress toward the next generation of MLLMs with human-like visual cognition abilities. MaRs-VQA is available at huggingface.co/datasets/IrohXu/VCog-Bench. The training code of Qwen2-VCog is available at github.com/IrohXu/Cognition-MLLM.

Bolin Lai, Xiaoliang Dai, Lawrence Chen et al.

Generating instructional images of human daily actions from an egocentric viewpoint serves as a key step towards efficient skill transfer. In this paper, we introduce a novel problem -- egocentric action frame generation. The goal is to synthesize an image depicting an action in the user's context (i.e., action frame) by conditioning on a user prompt and an input egocentric image. Notably, existing egocentric action datasets lack the detailed annotations that describe the execution of actions. Additionally, existing diffusion-based image manipulation models are sub-optimal in controlling the state transition of an action in egocentric image pixel space because of the domain gap. To this end, we propose to Learn EGOcentric (LEGO) action frame generation via visual instruction tuning. First, we introduce a prompt enhancement scheme to generate enriched action descriptions from a visual large language model (VLLM) by visual instruction tuning. Then we propose a novel method to leverage image and text embeddings from the VLLM as additional conditioning to improve the performance of a diffusion model. We validate our model on two egocentric datasets -- Ego4D and Epic-Kitchens. Our experiments show substantial improvement over prior image manipulation models in both quantitative and qualitative evaluation. We also conduct detailed ablation studies and analysis to provide insights in our method. More details of the dataset and code are available on the website (https://bolinlai.github.io/Lego_EgoActGen/).

Sangmin Lee, Bolin Lai, Fiona Ryan et al. · gatech

Understanding social interactions involving both verbal and non-verbal cues is essential for effectively interpreting social situations. However, most prior works on multimodal social cues focus predominantly on single-person behaviors or rely on holistic visual representations that are not aligned to utterances in multi-party environments. Consequently, they are limited in modeling the intricate dynamics of multi-party interactions. In this paper, we introduce three new challenging tasks to model the fine-grained dynamics between multiple people: speaking target identification, pronoun coreference resolution, and mentioned player prediction. We contribute extensive data annotations to curate these new challenges in social deduction game settings. Furthermore, we propose a novel multimodal baseline that leverages densely aligned language-visual representations by synchronizing visual features with their corresponding utterances. This facilitates concurrently capturing verbal and non-verbal cues pertinent to social reasoning. Experiments demonstrate the effectiveness of the proposed approach with densely aligned multimodal representations in modeling fine-grained social interactions. Project website: https://sangmin-git.github.io/projects/MMSI.

Zeyi Huang, Yuyang Ji, Xiaofang Wang et al.

Long-form video understanding with Large Vision Language Models is challenged by the need to analyze temporally dispersed yet spatially concentrated key moments within limited context windows. In this work, we introduce VideoMindPalace, a new framework inspired by the "Mind Palace", which organizes critical video moments into a topologically structured semantic graph. VideoMindPalace organizes key information through (i) hand-object tracking and interaction, (ii) clustered activity zones representing specific areas of recurring activities, and (iii) environment layout mapping, allowing natural language parsing by LLMs to provide grounded insights on spatio-temporal and 3D context. In addition, we propose the Video MindPalace Benchmark (VMB), to assess human-like reasoning, including spatial localization, temporal reasoning, and layout-aware sequential understanding. Evaluated on VMB and established video QA datasets, including EgoSchema, NExT-QA, IntentQA, and the Active Memories Benchmark, VideoMindPalace demonstrates notable gains in spatio-temporal coherence and human-aligned reasoning, advancing long-form video analysis capabilities in VLMs.

Bolin Lai, Felix Juefei-Xu, Miao Liu et al.

Text-guided image manipulation has experienced notable advancement in recent years. In order to mitigate linguistic ambiguity, few-shot learning with visual examples has been applied for instructions that are underrepresented in the training set, or difficult to describe purely in language. However, learning from visual prompts requires strong reasoning capability, which diffusion models are struggling with. To address this issue, we introduce a novel multi-modal autoregressive model, dubbed $\textbf{InstaManip}$, that can $\textbf{insta}$ntly learn a new image $\textbf{manip}$ulation operation from textual and visual guidance via in-context learning, and apply it to new query images. Specifically, we propose an innovative group self-attention mechanism to break down the in-context learning process into two separate stages -- learning and applying, which simplifies the complex problem into two easier tasks. We also introduce a relation regularization method to further disentangle image transformation features from irrelevant contents in exemplar images. Extensive experiments suggest that our method surpasses previous few-shot image manipulation models by a notable margin ($\geq$19% in human evaluation). We also find our model can be further boosted by increasing the number or diversity of exemplar images.

Bolin Lai, Sam Toyer, Tushar Nagarajan et al. · meta-ai

Predicting future human behavior is an increasingly popular topic in computer vision, driven by the interest in applications such as autonomous vehicles, digital assistants and human-robot interactions. The literature on behavior prediction spans various tasks, including action anticipation, activity forecasting, intent prediction, goal prediction, and so on. Our survey aims to tie together this fragmented literature, covering recent technical innovations as well as the development of new large-scale datasets for model training and evaluation. We also summarize the widely-used metrics for different tasks and provide a comprehensive performance comparison of existing approaches on eleven action anticipation datasets. This survey serves as not only a reference for contemporary methodologies in action anticipation, but also a guideline for future research direction of this evolving landscape.

Xinpeng Li, Bolin Lai, Hardy Chen et al.

We introduce Omni-MMSI, a new task that requires comprehensive social interaction understanding from raw audio, vision, and speech input. The task involves perceiving identity-attributed social cues (e.g., who is speaking what) and reasoning about the social interaction (e.g., whom the speaker refers to). This task is essential for developing AI assistants that can perceive and respond to human interactions. Unlike prior studies that operate on oracle-preprocessed social cues, Omni-MMSI reflects realistic scenarios where AI assistants must perceive and reason from raw data. However, existing pipelines and multi-modal LLMs perform poorly on Omni-MMSI because they lack reliable identity attribution capabilities, which leads to inaccurate social interaction understanding. To address this challenge, we propose Omni-MMSI-R, a reference-guided pipeline that produces identity-attributed social cues with tools and conducts chain-of-thought social reasoning. To facilitate this pipeline, we construct participant-level reference pairs and curate reasoning annotations on top of the existing datasets. Experiments demonstrate that Omni-MMSI-R outperforms advanced LLMs and counterparts on Omni-MMSI. Project page: https://sampson-lee.github.io/omni-mmsi-project-page.

Shiqi Peng, Bolin Lai, Guangyu Yao et al.

Vertebral body (VB) segmentation is an important preliminary step towards medical visual diagnosis for spinal diseases. However, most previous works require pixel/voxel-wise strong supervisions, which is expensive, tedious and time-consuming for experts to annotate. In this paper, we propose a Weakly supervised Iterative Spinal Segmentation (WISS) method leveraging only four corner landmark weak labels on a single sagittal slice to achieve automatic volumetric segmentation from CT images for VBs. WISS first segments VBs on an annotated sagittal slice in an iterative self-training manner. This self-training method alternates between training and refining labels in the training set. Then WISS proceeds to segment the whole VBs slice by slice with a slice-propagation method to obtain volumetric segmentations. We evaluate the performance of WISS on a private spinal metastases CT dataset and the public lumbar CT dataset. On the first dataset, WISS achieves distinct improvements with regard to two different backbones. For the second dataset, WISS achieves dice coefficients of $91.7\%$ and $83.7\%$ for mid-sagittal slices and 3D CT volumes, respectively, saving a lot of labeling costs and only sacrificing a little segmentation performance.

Jia Li, Wenjie Zhao, Shijian Deng et al.

Online egocentric gaze estimation predicts where a camera wearer is looking from first-person video using only past and current frames, a task essential for augmented reality and assistive technologies. Unlike third-person gaze estimation, this setting lacks explicit head or eye signals, requiring models to infer current visual attention from sparse, indirect cues such as hand-object interactions and salient scene content. We observe that gaze exhibits strong temporal continuity during goal-directed activities: knowing where a person looked recently provides a powerful prior for predicting where they look next. Inspired by vision-conditioned autoregressive decoding in vision-language models, we propose ARGaze, which reformulates gaze estimation as sequential prediction: at each timestep, a transformer decoder predicts current gaze by conditioning on (i) current visual features and (ii) a fixed-length Gaze Context Window of recent gaze target estimates. This design enforces causality and enables bounded-resource streaming inference. We achieve state-of-the-art performance across multiple egocentric benchmarks under online evaluation, with extensive ablations validating that autoregressive modeling with bounded gaze history is critical for robust prediction. We will release our source code and pre-trained models.

Bolin Lai, Xudong Wang, Saketh Rambhatla et al.

Latent diffusion has become the default paradigm for visual generation, yet we observe a persistent reconstruction-generation trade-off as latent dimensionality increases: higher-capacity autoencoders improve reconstruction fidelity but generation quality eventually declines. We trace this gap to the different behaviors in high-frequency encoding and decoding. Through controlled perturbations in both RGB and latent domains, we analyze encoder/decoder behaviors and find that decoders depend strongly on high-frequency latent components to recover details, whereas encoders under-represent high-frequency contents, yielding insufficient exposure and underfitting in high-frequency bands for diffusion model training. To address this issue, we introduce FreqWarm, a plug-and-play frequency warm-up curriculum that increases early-stage exposure to high-frequency latent signals during diffusion or flow-matching training -- without modifying or retraining the autoencoder. Applied across several high-dimensional autoencoders, FreqWarm consistently improves generation quality: decreasing gFID by 14.11 on Wan2.2-VAE, 6.13 on LTX-VAE, and 4.42 on DC-AE-f32, while remaining architecture-agnostic and compatible with diverse backbones. Our study shows that explicitly managing frequency exposure can successfully turn high-dimensional latent spaces into more diffusible targets.

Bolin Lai, Sangmin Lee, Xu Cao et al.

Text-image-to-video (TI2V) generation is a critical problem for controllable video generation using both semantic and visual conditions. Most existing methods typically add visual conditions to text-to-video (T2V) foundation models by finetuning, which is costly in resources and only limited to a few predefined conditioning settings. To tackle this issue, we introduce a unified formulation for TI2V generation with flexible visual conditioning. Furthermore, we propose an innovative training-free approach, dubbed FlexTI2V, that can condition T2V foundation models on an arbitrary amount of images at arbitrary positions. Specifically, we firstly invert the condition images to noisy representation in a latent space. Then, in the denoising process of T2V models, our method uses a novel random patch swapping strategy to incorporate visual features into video representations through local image patches. To balance creativity and fidelity, we use a dynamic control mechanism to adjust the strength of visual conditioning to each video frame. Extensive experiments validate that our method surpasses previous training-free image conditioning methods by a notable margin. We also show more insights of our method by detailed ablation study and analysis.

Wenqi Jia, Bolin Lai, Miao Liu et al.

Understanding and predicting human visuomotor coordination is crucial for applications in robotics, human-computer interaction, and assistive technologies. This work introduces a forecasting-based task for visuomotor modeling, where the goal is to predict head pose, gaze, and upper-body motion from egocentric visual and kinematic observations. We propose a \textit{Visuomotor Coordination Representation} (VCR) that learns structured temporal dependencies across these multimodal signals. We extend a diffusion-based motion modeling framework that integrates egocentric vision and kinematic sequences, enabling temporally coherent and accurate visuomotor predictions. Our approach is evaluated on the large-scale EgoExo4D dataset, demonstrating strong generalization across diverse real-world activities. Our results highlight the importance of multimodal integration in understanding visuomotor coordination, contributing to research in visuomotor learning and human behavior modeling.

Shiqi Peng, Bolin Lai, Guangyu Yao et al.

Spinal metastasis is the most common disease in bone metastasis and may cause pain, instability and neurological injuries. Early detection of spinal metastasis is critical for accurate staging and optimal treatment. The diagnosis is usually facilitated with Computed Tomography (CT) scans, which requires considerable efforts from well-trained radiologists. In this paper, we explore a learning-based automatic bone quality classification method for spinal metastasis based on CT images. We simultaneously take the posterolateral spine involvement classification task into account, and employ multi-task learning (MTL) technique to improve the performance. MTL acts as a form of inductive bias which helps the model generalize better on each task by sharing representations between related tasks. Based on the prior knowledge that the mixed type can be viewed as both blastic and lytic, we model the task of bone quality classification as two binary classification sub-tasks, i.e., whether blastic and whether lytic, and leverage a multiple layer perceptron to combine their predictions. In order to make the model more robust and generalize better, self-paced learning is adopted to gradually involve from easy to more complex samples into the training process. The proposed learning-based method is evaluated on a proprietary spinal metastasis CT dataset. At slice level, our method significantly outperforms an 121-layer DenseNet classifier in sensitivities by $+12.54\%$, $+7.23\%$ and $+29.06\%$ for blastic, mixed and lytic lesions, respectively, meanwhile $+12.33\%$, $+23.21\%$ and $+34.25\%$ at vertebrae level.

Bolin Lai, Fiona Ryan, Wenqi Jia et al.

Egocentric gaze anticipation serves as a key building block for the emerging capability of Augmented Reality. Notably, gaze behavior is driven by both visual cues and audio signals during daily activities. Motivated by this observation, we introduce the first model that leverages both the video and audio modalities for egocentric gaze anticipation. Specifically, we propose a Contrastive Spatial-Temporal Separable (CSTS) fusion approach that adopts two modules to separately capture audio-visual correlations in spatial and temporal dimensions, and applies a contrastive loss on the re-weighted audio-visual features from fusion modules for representation learning. We conduct extensive ablation studies and thorough analysis using two egocentric video datasets: Ego4D and Aria, to validate our model design. We demonstrate the audio improves the performance by +2.5% and +2.4% on the two datasets. Our model also outperforms the prior state-of-the-art methods by at least +1.9% and +1.6%. Moreover, we provide visualizations to show the gaze anticipation results and provide additional insights into audio-visual representation learning. The code and data split are available on our website (https://bolinlai.github.io/CSTS-EgoGazeAnticipation/).

Peng Wang, Yuhsuan Wu, Bolin Lai et al.

Objectives: to propose a fully-automatic computer-aided diagnosis (CAD) solution for liver lesion characterization, with uncertainty estimation. Methods: we enrolled 400 patients who had either liver resection or a biopsy and was diagnosed with either hepatocellular carcinoma (HCC), intrahepatic cholangiocarcinoma, or secondary metastasis, from 2006 to 2019. Each patient was scanned with T1WI, T2WI, T1WI venous phase (T2WI-V), T1WI arterial phase (T1WI-A), and DWI MRI sequences. We propose a fully-automatic deep CAD pipeline that localizes lesions from 3D MRI studies using key-slice parsing and provides a confidence measure for its diagnoses. We evaluate using five-fold cross validation and compare performance against three radiologists, including a senior hepatology radiologist, a junior hepatology radiologist and an abdominal radiologist. Results: the proposed CAD solution achieves a mean F1 score of 0.62, outperforming the abdominal radiologist (0.47), matching the junior hepatology radiologist (0.61), and underperforming the senior hepatology radiologist (0.68). The CAD system can informatively assess its diagnostic confidence, i.e., when only evaluating on the 70% most confident cases the mean f1 score and sensitivity at 80% specificity for HCC vs. others are boosted from 0.62 to 0.71 and 0.84 to 0.92, respectively. Conclusion: the proposed fully-automatic CAD solution can provide good diagnostic performance with informative confidence assessments in finding and discriminating liver lesions from MRI studies.

Xiao-Yun Zhou, Bolin Lai, Weijian Li et al.

Landmark localization plays an important role in medical image analysis. Learning based methods, including CNN and GCN, have demonstrated the state-of-the-art performance. However, most of these methods are fully-supervised and heavily rely on manual labeling of a large training dataset. In this paper, based on a fully-supervised graph-based method, DAG, we proposed a semi-supervised extension of it, termed few-shot DAG, \ie five-shot DAG. It first trains a DAG model on the labeled data and then fine-tunes the pre-trained model on the unlabeled data with a teacher-student SSL mechanism. In addition to the semi-supervised loss, we propose another loss using JS divergence to regulate the consistency of the intermediate feature maps. We extensively evaluated our method on pelvis, hand and chest landmark detection tasks. Our experiment results demonstrate consistent and significant improvements over previous methods.

Bolin Lai, Yuhsuan Wu, Xiao-Yun Zhou et al.

Lesion detection serves a critical role in early diagnosis and has been well explored in recent years due to methodological advancesand increased data availability. However, the high costs of annotations hinder the collection of large and completely labeled datasets, motivating semi-supervised detection approaches. In this paper, we introduce mean teacher hetero-modal detection (MTHD), which addresses two important gaps in current semi-supervised detection. First, it is not obvious how to enforce unlabeled consistency constraints across the very different outputs of various detectors, which has resulted in various compromises being used in the state of the art. Using an anchor-free framework, MTHD formulates a mean teacher approach without such compromises, enforcing consistency on the soft-output of object centers and size. Second, multi-sequence data is often critical, e.g., for abdominal lesion detection, but unlabeled data is often missing sequences. To deal with this, MTHD incorporates hetero-modal learning in its framework. Unlike prior art, MTHD is able to incorporate an expansive set of consistency constraints that include geometric transforms and random sequence combinations. We train and evaluate MTHD on liver lesion detection using the largest MR lesion dataset to date (1099 patients with >5000 volumes). MTHD surpasses the best fully-supervised and semi-supervised competitors by 10.1% and 3.5%, respectively, in average sensitivity.

Bolin Lai, Yuhsuan Wu, Xiaoyu Bai et al.

Using radiological scans to identify liver tumors is crucial for proper patient treatment. This is highly challenging, as top radiologists only achieve F1 scores of roughly 80% (hepatocellular carcinoma (HCC) vs. others) with only moderate inter-rater agreement, even when using multi-phase magnetic resonance (MR) imagery. Thus, there is great impetus for computer-aided diagnosis (CAD) solutions. A critical challenge is to robustly parse a 3D MR volume to localize diagnosable regions of interest (ROI), especially for edge cases. In this paper, we break down this problem using a key-slice parser (KSP), which emulates physician workflows by first identifying key slices and then localizing their corresponding key ROIs. To achieve robustness, the KSP also uses curve-parsing and detection confidence re-weighting. We evaluate our approach on the largest multi-phase MR liver lesion test dataset to date (430 biopsy-confirmed patients). Experiments demonstrate that our KSP can localize diagnosable ROIs with high reliability: 87% patients have an average 3D overlap of >= 40% with the ground truth compared to only 79% using the best tested detector. When coupled with a classifier, we achieve an HCC vs. others F1 score of 0.801, providing a fully-automated CAD performance comparable to top human physicians.